Molecular modeling techniques are widely used to discover drug candidates for selective disease. In the present study, ligand-based drug design techniques have been explored to find the structural requirement of diarylpropionitrile derivatives, a group of non-steroidal estrogen receptor (ER) modulators for selective binding to receptor subtypes. 2D/3D quantitative structure activity relationship (QSAR) and pharmacophore space modeling studies have been explored for this purpose. The classical QSAR models (Ra 2 = 0.870, Qa 2 = 0.813, Ra-pred 2 = 0.636; Rb 2 = 0.853, Qb 2 = 0.745, Rb-pred 2 = 0.565) show the importance of molecular refractivity, electronic contribution of atoms C3, C7, C13 and C14, and R2 and R4 substituents (Fig. 1) for specificity. The 3D QSAR, molecular fields (CoMFA, Ra 2 = 0.999, Qa 2 = 0.679, Ra-pred 2 = 0.678 and Rb 2 = 0.999, Qb 2 = 0.611, Rb-pred 2 = 0.691) and similarity (CoMSIA, Ra 2 = 0.999, Qa 2 = 0.670, Ra-pred 2 = 0.686 and Rb 2 = 0.999, Qb 2 = 0.671, Rb-pred 2 = 0.590) analyses show contour maps of steric, hydrophobic along with hydrogen bond (HB) donor and acceptor are important factors for binding affinity to both a- and b-subtypes. In addition, electronic contribution is crucial for a-subtype binding. Pharmacophore models derive the importance of HB acceptor and donor, aromatic ring, molecular steric, and hydrophobic interactions for selective binding to receptor subtypes. The derived models are correlated with structure-based molecular docking study, explaining the significant interactions between receptor and ligand for selective subtypes binding.