Classical conditioning is at the heart of most learning processes. It is thus essential that we develop accurate models of conditioning phenomena and data. In this paper we review the different uses of computational models in exploring conditioning, as simulators and as psychological models by proxy.