This is hard. Because I want to give you books where you could easily LEARN physics and learn about the problem solving methodology, but at the same time, learning science comes from DOING science and not from reading about science.

What I mean is this:

So you can read the Feynman Lectures but without having problem sets at the end, you're not going to retiain more than an abstract view of the situation.

On the other hand, I could also suggest Griffith's "Introduction to Electrodynamics" Which is widely regarded as the best resource for learning about Electrodynamics, but you'd be scared off by the time you got to the first chapter if you weren't prepared to do multivariable calculus.

In either of these cases though you still aren't learning about the science, you are learning about how to calculate the things we already know. So in this case I would suggest G. Polya's "How to Solve it" while focused on the pedagogy of mathematics, also offers a great way to learn about and solve problems.

Of course there are popular science books, and those are fine if you are fine with having an abstract view of the course of science

Six Easy Pieces - Feynman A Brief History of Time - Hawking Cosmos - Sagan Demon Haunted World - Sagan What Evolution Is - Ernst Mayr

For me though the more satisfying route in the long run is to sit down and look at the problems others have solved, and try to solve them yourself. See what kind of conclusions you come to.