Run-to-run optimization exploits the repetitive nature of batch processes to adapt the operating policy in the presence of uncertainty. For problems where terminal constraints play a dominant role in the optimization, the system can be operated close to the optimum simply by satisfying terminal constraints. The input is parameterized by using the knowledge of the shape of the optimal solution and, in the presence of uncertainty, the input parameters are adapted to meet the terminal constraints. When the number of input parameters is greater than the number of terminal constraints, an adaptation methodology based on a projection matrix derived from the gain matrix between the input parameters and the terminal constraints is proposed. The run-to-run optimization scheme is illustrated in simulation for the minimization of the batch time of an emulsion polymerization process with terminal constraints on conversion and number average molecular weight.