In this study, the application of a repeated process estimation-optimization strategy is investigated for a methanol synthesis reactor system in the presence of a slowly deactivating catalyst. The purpose of this strategy is to update the optimal transition of the control variable profile with feedback from the process measurements. A state and parameter estimation method is formulated where the objective is to determine the current reactor state and model parameter from available process measurements data. In this strategy, the corresponding optimization problems are formulated as DAE-constrained optimization problems that are solved by using full discretization and large-scale NLP solver. Simulation results indicate that the strategy is able to track the theoretical optimum profile of the selected control variable as the catalyst deactivates. Moreover, with the formulated strategy, the performance of the reactor system in the presence of a long-term catalyst deactivation and unexpected plant operational changes can be improved significantly. (C) 2012 Elsevier Ltd. All rights reserved.