A screening method for selective polymorphic crystal growth was proposed using molecular modeling as an assistant tool in this work. Three typical pharmaceutical polymorphic crystals which are theophylline, L-serine and fumaric acid were chosen as case studies to investigate their selective growth on self-assembled monolayer (SAM) prepared by depositing 3-mercaptopropionic acid (MPA) on gold. Molecular modeling was applied to analyze the arrangement of MPA molecules on the surface, as well as to value the binding energy between these three crystal faces with MPA SAM. The simulation results were used to predict the effect of SAM on the two dimensional selective growth of the polymorphic crystals, and crystallization experiments were carried out to verify the modeling prediction. The results indicate that MPA SAM can directionally induce the nucleation of polymorphic crystals by geometric and chemical interactions matching. Molecular modeling can be as an effective assistant tool to reveal the matching degree, which can give valuable suggestions on the polymorphic drug design.