In this article we address the production planning of parallel multiproduct batch reactors with sequence-dependent changeovers, a challenging problem that has been motivated by a real-world application of a specialty chemicals business. We propose two production planning models that anticipate the impact of the changeovers in this batch processing problem. The first model is based on underestimating the effects of the changeovers that leads to an MILP problem of moderate size. The second model incorporates sequencing constraints that yield very accurate predictions, but at the expense of a larger MILP problem. To solve large scale problems in terms of number of products and reactors, or length of the time horizon, we propose a decomposition technique based on rolling horizon scheme and also a relaxation of the detailed planning model. Several examples are presented to illustrate the performance of the proposed models. (C) 2007 American Institute of Chemical Engineers.