You can only make educated guesses at this - your revenue will be tied directly to the number of copies of the game you sell, and if there were a way to compute that with any accuracy, we'd all be much richer.
Most companies (it's usually publishers, specifically, that do this) do what is essentially market research. They mine data from as many sources as they can, public and private, to get any idea of how well games with similar features did on similar platforms at similar times of the year. If a company has released a similar product themselves, they'll have that specific market data to draw upon, as well.
Today, the indie boom has made this a little easier to do, because there's more publically-available data floating around. SteamSpy, for example, while not perfect, is much better than anything that was available to the small-time developer twenty years ago. Many indies are much more open about their successes (or failures) post-launch, as well, and you can often find postmortems by developers that detail some sales information. For example, the developer of Cogmind (not at all similar to your game) published recaps of year one and year two of the life of his game. If you search around, you may be able to find similar postmortems from developers of games like yours.
Outside of correlative data like the above, you can use more direct numbers from other ways users have engaged with your game before release (if you have them). For example, does your game have a Twitter account? How many followers does it have? How many mailing-list sign-ups do you have? YouTube views or subscriptions to your developer logs? Et cetera. That kind of pre-ship engagement traffic can help give you rough estimates of the order of magnitude (at least) of what your sales could be. Provided you don't turn everybody off with a bad price point, of course.