An approach is proposed for the robust identification and control of batch and semibatch processes. The batch experiments used for model identification are designed by minimizing the magnitude of the parameter uncertainties, and the effect of these uncertainties on the product quality achievable by optimal control is used as a stopping criterion for the identification procedure. The optimal control approach incorporates a quantification of the impact of both parameter and control implementation uncertainties on the performance of the optimal control policy. The approach is applied to the nucleation and growth of crystals with multiple characteristic dimensions, where the nominal parameters used in the simulation study are quantified from experimental data. (C) 2003 Elsevier Science Ltd. All rights reserved.