The cytochrome P450s are a superfamily of hemoprotein enzymes responsible for the metabolism of a wide variety of xenobiotic and endogenous compounds. The individual P450s exhibit unique substrate specificity and stereoselectivity profiles which reflect corresponding differences in primary sequence and tertiary structure. In the absence of an experimental structure models for mammalian P450s have been generated by their homology with bacterial P450s of known structure. The rather low sequence identity between target and template proteins renders P450 modeling a challenging task. However, the substrate recognition properties of several P450s are consistent with recently developed working models. This review summarizes the major concepts and current approaches of molecular modeling of P450s.