Travel guide · BOOKING
The Bella-Ruta Booking Timeline: What to Book, and Exactly When
Every rail operator's advance booking window, what the airfare research actually says, and the order to book things in.
- Best for
- Anyone with a route settled and no idea what to do first.
- Not ideal for
- Anyone who has not settled the route yet. Do that first — it is free.
- Last verified
- 20 August 2026
The rule underneath all of it
Book in order of how hard it is to change.
That is the whole method. Everything below is detail.
1. Long-haul flights. Biggest cost, least flexible, and they lock your dates.
2. Anything on a fixed timetable or that sells out. Trains, ferries, regional flights, anything tied to an event.
3. Accommodation. Usually movable.
4. Nothing. Genuinely nothing. Do not book activities for day nine of a trip you have not taken.
The mistake is rarely booking too late. It is booking in the wrong order — locking a flight in January and building a route around it in April.
Why flights and rail have different windows
Airlines load schedules eleven to twelve months out and sell from the moment they do.
Railways load schedules at the timetable change — in Europe, twice a year, in June and December. Some sell a year ahead. Some sell four weeks ahead. Some, like Renfe, behave differently from one month to the next.
So you cannot book a trip in one sitting. You book the flights, then you wait, then you book the trains. Diarising the second date is the part people forget.
The rail booking windows
Europe
| **Operator** | **Advance window** | **Notes** |
|---|---|---|
| Deutsche Bahn | 12 months | Official. Sparpreis from €21.99, Super Sparpreis from €6.99. Free cancellation of digital tickets within 3 hours of purchase |
| Eurostar | 10–11 months | Official |
| Amtrak | 11 months | Official. No fixed advance-purchase discount tier |
| SNCF (TGV INOUI, OUIGO Grande Vitesse) | 6 months | Changed 11 March 2026 — see below |
| Nightjet (ÖBB) | 180 days | Official. Shorter for dates after the mid-December timetable change |
| OUIGO Train Classique | 60 days | Official |
| Trenitalia | ~4 months, varies | Not officially published. Can shrink to ~30 days after a timetable change |
| Renfe | 15 days to 11 months | Erratic. Competitive routes up to 11 months, monopoly routes as little as 30 days |
Three things worth pulling out.
SNCF changed in March 2026 and most guides have not caught up. All TGV services operated by SNCF Voyageurs now go on sale six months before travel, replacing the old fixed batch release dates. SNCF Connect's own help page still says four months. If you read four months anywhere, it is stale.
Eurostar's connecting-service windows are the buried detail. You can book the Eurostar leg eleven months out, but the connecting TGV only 120 days out, the connecting German ICE only 90 days, Belgian services 4 to 8 months, and UK National Rail about 84 days. A London-to-Munich itinerary therefore takes three separate booking sessions.
Renfe will make you think a train does not exist. If you go online and do not see the service you expect with over a month to go, wait — the trains are probably not loaded yet.
Asia
| **Operator** | **Advance window** | **Notes** |
|---|---|---|
| Japan — smartEX (Tokaido/Sanyo/Kyushu Shinkansen) | 1 year, from 05:30 JST | Online only |
| Japan — everything else | 1 month, from 10:00 JST on the same calendar day of the preceding month | Consistent across JR Central, JR West and JR East |
| Thailand (SRT) | 90 days | Cut from 180 days in late April 2026 |
| Vietnam Railways | 60 days | For long-distance journeys |
| South Korea (Korail) | 30 days, from 07:00 KST | Sales close 20 minutes before departure |
| Taiwan (THSR) | 29 days | Extends to the following Sunday if the 29-day mark falls on a Friday or Saturday |
| Laos–China Railway | 3 to 7 days | Sources conflict. Effectively "book on the ground" |
The Asian pattern is the opposite of Europe's. Windows are short — a month or less almost everywhere except smartEX. That is good news: you cannot get it wrong by being too early, and you only need to be organised for one date.
Diarise the exact hour. 10:00 Japan time. 07:00 Korean time. These are real deadlines on popular services.
The airline side
How far ahead schedules open
| **Carrier** | **Window** |
|---|---|
| Iberia | 365 days |
| Korean Air | 362 |
| Qatar Airways | 361 |
| Cathay Pacific, Lufthansa | 360 |
| British Airways | 355 |
| Qantas | 354 |
| Emirates | 340 |
| Delta | 331 |
| American, Alaska, Japan Airlines | 330 |
| Air France, KLM, Etihad, Virgin Atlantic | 11 months |
| JetBlue | 6–8 months |
| Ryanair | 5–6 months |
Safe statement: 330 to 360 days, roughly eleven to twelve months. US carriers cluster at 330 to 331. Gulf, Asian and European flag carriers push to 355 to 365.
When to actually book — and why the headline number is misleading
Two credible studies disagree, and the disagreement is instructive.
Expedia's 2026 Air Hacks report, published 17 February 2026, based on Expedia bookings from December 2024 to November 2025:
- International economy: cheapest at 8 to 15 days ahead, average \$585
- A strong secondary window at 31 to 45 days, saving about \$190 against booking 180+ days out
- Domestic economy: optimal at 15 to 30 days
- Friday is the cheapest day to depart (about 8% under Sunday) and the cheapest day to book (about 3%)
- Tuesday is cheapest for domestic departures, about 14% under Sunday
- August is about 29% cheaper than December, roughly \$120 a ticket
Google's own research, published 31 August 2022 and based on five years of airfare data:
- Europe: book 50 to 179 days ahead. Prices bottom out around 129 days
- US domestic: lowest 21 to 60 days out, bottoming at 44
- Thanksgiving: 36 to 74 days, lowest at 52. Christmas: 20 to 88, lowest at 22
- Midweek flights about 12% cheaper than weekend departures
- Nonstop fares about 20% higher than connecting
Do not act on the 8-to-15-day finding
It is almost certainly a selection effect rather than causal advice.
Last-minute international bookings skew towards flexible travellers taking whatever route happens to be cheap, and towards low-demand departures. The people booking eight days out are not the same people as those booking six months out, and they are not buying the same trips.
If you have fixed leave dates and a specific two-week route, waiting until eight days out is not a strategy. It is a gamble with your annual leave as the stake.
The reconciliation
Google measured price minima across five years of route data. Expedia measured what its own bookers paid. Both are credible and they are measuring different things.
For a working professional booking a fixed multi-city trip: two to four months ahead for long-haul international.
That sits inside Google's 50-to-179-day range for Europe, near its 129-day low point, and it overlaps Expedia's defensible 31-to-45-day secondary window. It is also far enough out that you are not gambling.
The countdown
6 to 12 months out
Settle the route. Cities, order, nights, connections, entry and exit airports. This costs nothing and everything else depends on it.
Check whether any leg of your route has a limited-frequency flight — the four-days-a-week Chiang Mai to Luang Prabang service is the example that reshapes a whole itinerary. Find those before you fix anything.
3 to 4 months out
Book the long-haul flights, as a multi-city or open-jaw booking if the route is a line rather than a loop.
Book accommodation in the tightest-supply city — anywhere with a fixed event, and Kyoto in autumn.
Book anything with a hard capacity limit. A Ha Long Bay cruise. A specific sleeper compartment. The Pena Palace timed slot in Sintra.
2 to 3 months out
European rail, where the window is open. DB opens 12 months out, Eurostar 10 to 11, SNCF 6, Nightjet 6.
Apply for visas with a buffer. Vietnam e-visa: 3 to 5 working days. Laos eVisa: about 3 working days, and it can only be applied for between 5 and 60 days before arrival.
1 month out
Japanese rail, at 10:00 Japan time on the same calendar day of the preceding month.
Korean rail, 30 days out at 07:00 Korean time.
Thai rail, within the 90-day window.
29 days out
Taiwan HSR. Book all legs at once and take the Early Bird discount, which requires booking at least five days ahead and is only available online, in the app, or at partner convenience stores.
3 days out
The arrival cards. These are new, they are free, and they are the most common thing people miss in 2026:
- Thailand — TDAC, within 3 days, tdac.immigration.go.th
- Taiwan — Taiwan Arrival Card, twac.immigration.gov.tw, mandatory since 1 October 2025
- South Korea — e-Arrival Card, within 3 days, mandatory since 1 January 2026 for anyone not holding a K-ETA
- Vietnam — pre-arrival declaration for air arrivals, no earlier than 3 days before
On arrival
Nothing. You should have nothing left to book.
Worked examples
Japan, 14 days
| **When** | **What** |
|---|---|
| 6 months | Route settled. Tokyo in, Osaka out |
| 3–4 months | Flights, as open-jaw. Kyoto accommodation |
| 1 month, 10:00 JST | Shinkansen reserved seats |
| Any time | JR Pass, bought on japanrailpass.net — only the official site lets you reserve seats online before you land |
Korea, 10 days
| **When** | **What** |
|---|---|
| 3–4 months | Flights. Decide ICN round trip versus ICN-in / PUS-out |
| 30 days, 07:00 KST | KTX tickets on korail.com, as a non-member if the membership flow demands a Korean phone number |
| 3 days | e-Arrival Card |
Europe, a two-city long weekend
| **When** | **What** |
|---|---|
| 3–4 months | Flights. Check the Sunday evening return exists before booking the outbound |
| 3–6 months | Rail, depending on operator. DB 12 months, SNCF 6, Trenitalia ~4, Renfe when it feels like it |
| At booking | Any timed-entry attraction |
Price monitoring, briefly
Set alerts on the routes you are watching, and set them the day the route is settled rather than the day you decide to book.
Two things worth knowing from the research: nonstop fares run about 20% above connecting fares, and midweek departures run about 12% below weekend ones. Both are worth checking against what a connection or a shifted departure day does to your usable day count.
A cheaper flight that lands at 22:00 has cost you a day, and on most trips a day is worth more than the fare difference.
Common mistakes
Booking flights before the route is settled. You end up building the trip around a fare instead of a map.
Booking accommodation before inter-city transport. Rooms are flexible. Timetables are not.
Trying to book everything in one sitting. Impossible. The windows do not line up.
Reading a stale booking window. SNCF changed in March 2026, Thailand changed in April 2026. A lot of published advice is wrong on both.
Buying a JR Pass through an agency. You lose online seat reservation.
Missing the arrival cards. Four of the countries in this content set now require one, all introduced between May 2025 and January 2026.
Waiting until 8 to 15 days out because a headline said fares are lowest then.
What needs checking before this is published
1. Renfe's official advance booking window. Renfe does not publish one and its own FAQ pages returned 404s. Verify on renfe.com for the specific route.
2. Trenitalia's official window. Not stated on Trenitalia's own site. Seat61 says about 4 months, Trainline says 6. Unresolved.
3. Trenitalia Super Economy's "midnight of the 5th day before departure" deadline. Only found on a reseller site. Not stated in the article.
4. SNCF's current state. The official March 2026 six-month announcement and SNCF Connect's own FAQ contradict each other. Test empirically by trying to book a TGV five months out.
5. Airline schedule-loading windows come from a reseller aggregation, not per-airline official confirmation.
6. A post-2022 Google Flights booking-window study. The August 2022 analysis appears to be Google's most recent published research. Flag the date whenever it is cited.
7. Hopper. The only verifiable Hopper guidance is from October 2017. It is not cited here and should not be.
8. Laos–China Railway window. Three days versus seven days, unresolved.
9. SRT 90-day effective date. 24 versus 30 April 2026.
10. Anny's call: this article contains the strongest single piece of contrarian, defensible advice in the set — that the widely-quoted "book 8 to 15 days out" finding is a selection artifact and should not be followed by anyone with fixed leave. Worth its own Reel.