The relationship between the molecular structures of lubricant additives and their antiwear properties was assessed using 36 nitrogen-containing heterocyclic organic compounds to generate quantitative structure tribo-ability relationship (QSTR) models. In the modeling process, the molecular structure of the base oil was employed as the solvation factor, employing n-octadecane as a substitute for the liquid paraffin used to generate the original experimental data. Comparative molecular field analysis (CoMFA)-QSTR and comparative molecular similarity indices analysis (CoMSIA)-QSTR models incorporating solvation by n-octadecane were established and compared with QSTR models that did not involve solvation. The solvation CoMFA-QSTR and CoMSIA-QSTR models exhibited excellent fitting abilities and were highly robust when predicting lubricant performance. These models were superior to nonsolvation models in this regard. These models also generated useful information regarding potential changes to molecular structure to improve antiwear properties. Electrostatic fields were found to be very important factors in antiwear models. Because the original experiments used a nonpolar base oil, the solvation factor had no obvious effect on the antiwear characteristics of the polar additives and so the predictive abilities of the solvation and nonsolvation models were similar.