The best approach will depend a lot on your data. A good first step might be to look at pairs of parameters using cftool. This will open the curve-fitting tool. When you specify a dependent (x) variable and independent or result value (y), it will give you a variety of functional forms and optimize the fit to the data. This can be very quick and easy, but the pre-defined functions are not always good at describing the data. But I would begin exploring the relationships there.
Sorry the answer is so vague, but multivariate analysis is a big subject, so without more to go on this is about all I can say.