ObjectiveTo construct the three dimensional pharmacophore models of inhibitors of vascular endothelial growth factor receptor 2(VEGFR-2), so as to provide a theoretical model for designing novel VEGFR-2 inhibitors. MethodsThe model was derived from 30 reported VEGFR-2 inhibitors (21 molecules as training set and the rest as testing set)with structural variety and resembling binding mode, whose IC50 values spanned 5 magnitudes by using Catalyst Software in Discovery Studio 2.5 Package. ResultsThe best model consists of one hydrogen bond receptor, two hydrophobic cores, one aromatic plane and 4 excluded volumes. The predictive relativity (R) was 0.89. ConclusionCross-validation results indicate that the constructed model has satisfactory predictivity and can be used to screen molecule databases for searching novel lead compounds which can inhibit VEGFR-2.