The optimization of emulsion polymerization processes can benefit from the application of neural networks (NNs). The aim of this paper is to present yet another important example of how NNs can help with the optimization of emulsion polymerization reactors, and more specifically, with the "inverse" problem, i.e., starting from the polymer's properties and productivity to determine the reactor's operating conditions. The procedure presented in this paper combines a mathematical model with a NN that predicts appropriate operating conditions for the reactor with a maximum error of 5%. The specific example is with the emulsion polymerization of vinyl acetate. (C) 2004 Elsevier Ltd. All rights reserved.