learning logistic regression

If you're really good at multiple linear/non-linear regression analysis, would it be hard to learn how to do a logistic regression?

Is that a lot different (different assumptions & tests) or is it similar enough to the 'normal' regression model that one would only have to learn a few more things?

Not sure what you mean by "really good", but if you know SPSS, you'll have a handle on binomial and multinomial logistic regressions in no time. Analysing model strength is not as straightforward, but everything else is pretty easy.