In this study, a QSAR model of neuraminidase (NA) type 1 (N1) was elevated. This map contained two hydrogen bond acceptor features, one hydrogen bond donor features, and one positive ionizable feature. In the second step, we created the interaction maps in the active sites on the neuraminidase type2, and type7 (N2 and N7) protein structures. The structure-based pharmacophore map was showed the features on every amino acid in the active site on the protein structure. The third step was pharmacophore comparison, root-mean-squared error (RMSE) was reported for the matching pharmacophore features. The result showed that the maps of N1, N2, and N7 had subtle differences in distances of each features. We created the combined map for N1, N2, and N7 to resolving the difference in the three NA types. The combined map was employed to NCI database screening, then, the potent versatile inhibitors were elevated in the results.