A methodology for the optimization of semibatch reactors using dynamic programming is proposed. This includes synthesis of a mathematical model, analysis of the performance of the process at its present state, definition of a set of decision variables, and optimization and simplification of this optimum toward feasibility. The methodology was applied to an industrial case study in the fine chemical industry using the lowest product cost as the objective function. Stages of equal feed volume were used instead of stages of equal time length. The obtained multistep feed profile was transformed into a feasible two-step procedure, which resulted in similar performances. Both procedures resulted in increased cycle times, but this loss in productivity was more than compensated for by a higher yield.