This paper considers the model-based control of composition and thickness for a thin-film drying process used in the continuous manufacturing of pharmaceutical tablets. In this nonlinear distributed dynamical system, a drug formulation solution is coated onto a moving surface and then dried to form thin films of approximately 250 pm in thickness. A dynamic optimizer is designed that employs a first-principles process model to simulate the spatial distribution of solvent concentration in the film and the thin film shrinkage during drying. A critical parameter to describe the highly nonlinear dynamics of the thin-film drying is the mutual polymer solvent diffusion coefficient, which strongly depends on solvent concentration and film temperature. Two optimal control problems are studied for set point tracking of solvent concentration and minimization of energy consumption in the dryer while satisfying various operational and product quality constraints. An unscented Kalman filter is designed to facilitate the output feedback implementation of the dynamic optimizer and to estimate unmeasured thin-film quality attributes such as the film thickness. The performance of the model-based controller is compared to that of a proportional integral controller in two simulation case studies. The nonlinear model-based control strategy has improved versatility and the potential to reduce production of off spec material.