The reasonable prediction of metabolic clearance rate from the humanized in vitro system is valuable in drug discovery, which is commonly used in the identification and optimization of compounds that mostly like to process appropriate pharmacokinetic characteristics in humans. A detailed development of the general theory and, models underlying the prediction of in vivo hepatic drug metabolism from in vitro data were presented. Furthermore, the accuracy when extrapolating from in vitro data considering the in vitro-in vivo correlation, method-logical issues and potential solutions were discussed as well. This review can give us a better insight into exploring methods whereby human clearance can be accurately predicted from in vitro data in the process of new drug development.