You look at the world around you, and you take it apart into all its components. Then you take some of those components, throw them away, and plug in different ones, start it up and see what happens.
Stephen Hawking said he spent most of his first couple of years at Cambridge reading science fiction (and I believe that, because his grades weren't all that great).
I did that for 40 years or more. I never had any writer's block. I got up in the morning, sat down at the typewriter - now, computer - lit up a cigarette.
If you don't care about science enough to be interested in it on its own, you shouldn't try to write hard science fiction. You can write like Ray Bradbury and Harlan Ellison as much as you want.
A large fraction of the most interesting scientists have read a lot of SF at one time or another, either early enough that it may have played a part in their becoming scientists or at some later date just because they liked the ideas.
I don't think the scientific method and the science fictional method are really analogous. The thing about them is that neither is really practiced very much, at least not consciously. But the fact that they are methodical does relate them.
My old English buddy, John Rackham, wrote and told me what made science fiction different from all other kinds of literature - science fiction is written according to the science fiction method.
Stories where the author has known very little, but run a computer program that tells him how to construct a planet, and looked up specific things about rocketry and so on, really suck.