There's no easy answer. Even the top players can find it difficult to decide. Maybe a few considerations to help you decide...
- are any of rooks performing a more important role in their current position, compared to the other? Or likely to be doing so soon? If so, move the rook which is doing the least.
For example, if having castled kingside, is f2 vunerable to attack? Maybe moving the f1 rook will weaken your kingside too much. Or alternatively, maybe there's possibilities of playing f4 and wanting its advance to be supported by the rook.
- is it useful to vacate a particular square?
Sometimes the f1 rook will move to allow another piece to occupy it. Note: rook on f1 covers f2; bishop would cover g2; and knight covers h2. So may depend on where you think the opponent will attack and how best to defend. Also, sometimes the king may want f1 as an escape!
- coordination of the pieces
Sometimes you want to drop a bishop back to, say, b1 or c1. A rook isolated on a1 may not be happy about this. So maybe you want to move this rook to a central file before the bishop blocks the first rank.
- subject to attack
Are any of the rooks more likely to come under attack on their current square? e.g. the rook on a1 from a Black fianchetto bishop on g7? If so, you may want to move that rook off of the diagonal
- think carefully about open files
Not all open files are of equal value. It's been said that the file is only of use if you can infiltrate the opponent's position (e.g. 7th rank) using it. Sometimes an open file will offer no advantage if you control it.
Likewise, you need to consider the possibility of new files opening. It's common for GMs to play rooks to a closed or semi-open file either because they intend to open it more, or to prepare for their opponent doing so.