It seems to me that math is much more "formal" in the development of a theory and in proving it. Granted, that may be because I'm so fond of math - but still, science does not seem anywhere near as formal (except of course when the science is coupled tightly with math.)
Just to elaborate on your point, Mathematical "theories" are so much more formal because we rely upon logic, which in turn requires great intellectual rigor, to deduce the conclusions from very precisely formulated axioms. It is because of the formality and precision that we are able to "prove" our conclusions in Mathematics, while we never can do so in the same sense in scientific theories.
Scientific "theories", OTOH, are accepted (not proven) based upon observational results and experimental evidence (especially repeated failed attempts to falsify it), which are always incomplete, leaving open the door for future falsification.
While there are some differences between Mathematical and scientific "theories," it bears repeating that they still share a great many common attributes. In particular, both are conceptual frameworks that have broad explanatory power over the phenomona (or topics) within their respective scopes, and both must be conceptually falsifiable, the difference being that while a scientific theory is always at risk of being falsified, a properly deduced Mathematical theory (one with sound axioms and valid logical deductions), though capable of falsification, will never be, because our "proof tools" allow us to exclude the existence of any counter example.