3-D QSAR studies were conducted using a series of 39 compounds obtained in our laboratory to inhibit aminopeptidase N/CD13 (APN) for developing highly potent antitumor agents. Among constructed 3-D QSAR models, the best q 2 is 0.664 and the best r 2 is 0.995. Henceforth, the best 3-D QSAR model was applied for further chemical modification and optimization. Therefore, seven molecules were designed, and their activity values were predicted by the generated model. Their binding modes were elucidated by docking. Finally, both the high predicted activity values of the seven designed molecules and the favorable binding patterns emphasize the significance for further synthesis of these designed compounds.