Abstract: AIM To develop a three dimensional quantitative structure activity relationship (3D-QSAR) model and gain further insights into the requirements for potential P450_17α inhibitors. METHODS AND RESULTS A predictive 3D pharmacophore model was established based on comparative molecular field analysis (CoMFA). The correlation between the activities and structures was significant with cross validated value (R²_cv), non cross validated value (R²) and standard error of estimate (SEE) of 0.538, 0.799 and 0.257, respectively. According to this model, the predicted inhibition activities of three compounds synthesized in our laboratory were compatible to actual activities. CONCLUSION This model would contribute to the understanding of the interaction between the inhibitors and P450_17α and rational design of novel lead molecules.