Abstract We have devised a novel diffuse interface formulation to model the development of chemical and physical inhomogeneities, i.e. microstructure, during the process of casting drug-eluting coatings. These inhomogeneities, which depend on the coating constituents and manufacturing conditions, can have a profound affect on the rate and extent of drug release, and therefore the ability of coated medical devices to function successfully. By deriving the model equations in a time-dependent reference frame, we find that it is computationally viable to probe a wide, physically relevant range of material and process quantities. To illustrate the application of the model, we have evaluated the impact of manufacturing solvent, coating thickness and evaporation rate on microstructure development. Our results suggest that modifying these process conditions can have a strong and nearly discontinuous effect on coating microstructure, and therefore on drug release. Further, we demonstrate that the model can be applied to processes that involve the incremental application of the coating in layers or passes. This new model formulation, which can also be used to predict the kinetics of drug release, provides a tool to elucidate and quantify the relationships between process variables, microstructure and performance. Establishing these relationships can reduce empiricism in materials selection and process design, providing a facile and efficient means to tailor the underlying microstructure and achieve a desired drug-release behavior.