CYP3A4 is the main human metabolizing enzyme, and many clinically relevant drug/herb-drug interactions (DDIs/HDIs) involving CYP3A4 are due to mechanism-based inhibition. In this study, pharmacophore model together with molecular docking (MD) are used to rapidly screen the potential CYP3A4 mechanism-based inhibitors from Tripterygium wilfordii, and in vitro experiments are conducted to validate the computational data. The results showed that the rate of computational prediction could be improved based on a combination of pharmacophore model and MD, and a combination of computational approaches might be a useful tool to identify potential mechanism-based inhibitor of CYP3A4 from herbal medicines.