I wouldn't consider myself an experienced guy, but from what I have learned by listening to classical music of all periods and attempting to find out why the things work that I feel "just work", this is my impression for working with thematic material:
1) Repetition is not bad if it serves a purpose. I used to really avoid repeating my material in my writing because similar to
@Nova I was afraid of getting repetitive and losing the interest of the listener. But now I feel that it can be very important to repeat thematic material, especially if you plan to do smart and interesting things with it later, because the "new" things you are going to introduce when you develop things later in your piece are only going to work for the listener if they had enough time to digest the main ingredients first. I think this is why music of the classical period usually repeated the exposition part in live performance, because without recordings concertgoers usually had very few chances to hear a piece at all, and since you had their undivided attention in the concert hall, you could really hammer your themes in their brains before you blew their minds with your smart development of the material.
In the recording age, there is of course the possibility of anyone tuning out at any given moment, so the fear of losing them with repetition is real. But I think if you do repetitions and reprises in an original and interesting way that always adds or alters a little something while still feeding your listeners the important stuff for later on, it still works.
2) Different sections don't have to stay on their own. You can plan beforehand how they are going to interact. To me, a piece is always more interesting if at some point you realize as a listener that elements that previously stood on their own suddenly work beautifully together. This could be very simple, for example a few notes or a rhythmic pattern from your A theme continuing to play on when your B theme is introduced to form a simple rhythmic or ostinato accompaniment. It won't be in the foreground and won't steal the attention from the B theme, but it will create a satisfying unity between your material. What I especially like is if the different themes work in counterpoint, so you can suddenly have them play together at some later point in the piece for a real listening highlight. But there are so many different ways in which one can create a strong whole out of very different material.
3) This is maybe the most important point for me and not that different from the previous two: Thinking big in terms of structure is a good thing, but to me the smaller context is more important. I always find myself abandoning structural plans I had beforehand because in the process of writing I realize that the material has to grow organically in some other way. Because in the end, your listeners will judge the piece on the basis of what they are hearing and will quickly recognize if it offers a satisfying development musically. They are not going to have a checklist with structural elements of music and say "Okay, but here I was expecting the start of part 2 of the B section, what happened!? Boooo!"
I think of it in terms of a staircase. It is of course not unimportant if the stairs make a right turn or a left turn on their way upstairs, but nobody will care about that if your steps aren't even and walkable.
It is kind of a similar point to the one South Thames was making earlier. Intuition helps with that a lot, and when in doubt, I think it is always better to go with what feels right musically than to get tied in some structural corset.
Hope those ramblings made some sense.