Abstract: OBJECTIVE To establish a model of the relationship between the structure and activity of EGFR inhibitors, and to design new inhibitor molecules and predict their activity based on the information obtained on important structural factors affecting activity, so as to provide a basis for the design of inhibitor molecules. METHODS Discovery Studio 2019 software was used for 3D-QSAR research and partial least squares calculation. Autodock was used for molecular docking. Study two-dimensional interactions by using LigPlot. RESULTS The model had a high q²(0.521), and r²(r²_training=0.993, r²_test=0.916, r²_blind=0.940), indicated that the model had high predictive ability and fitting ability. CONCLUSION The prediction results show that the newly designed compounds have higher activity, providing reference for the design of EGFR inhibitor molecules.