Current industrial practice for control of primary nucleation (nucleation from a system without pre-existing crystalline matter) during crystallization from solution involves control of supersaturation generation, impurity levels, and solvent composition. Nucleation behavior remains largely unpredictable, however, due to the presence of container surfaces, dust, dirt, and other impurities that can provide heterogeneous nucleation sites, thus making the control and scale-up of processes that depend on primary nucleation difficult. To develop a basis for the rational design of surfaces to control nucleation during crystallization from solution, we studied the role of surface chemistry and morphology of various polymeric substrates on heterogeneous nucleation using aspirin as a model compound. Nucleation induction time statistics were utilized to investigate and quantify systematically the effectiveness of polymer substrates in inducing nucleation. The nucleation induction time study revealed that poly (4-acryloylmorpholine) and poly(2-carboxyethyl acrylate), each cross-linked by divinylbenzene, significantly lowered the nucleation induction time of aspirin while the other polymers were essentially inactive. In addition, we found the presence of nanoscopic pores on certain polymer surfaces led to order-of-magnitude faster aspirin nucleation rates when compared with surfaces without pores. We studied the preferred orientation of aspirin crystals on polymer films and found the nucleation-active polymer surfaces preferentially nucleated the polar facets of aspirin, guided by hydrogen bonds. A model based on interfacial free energies was also developed which predicted the same trend of polymer surface nucleation activities as indicated by the nucleation induction times.