The incremental addition of chain transfer agent (CTA) in an emulsion styrene-butadiene polymerization reaction was optimized, using a chemometric approach that involves (a) an experimental design, in which the amount of CTA in each addition were the factors; (b) an artificial neural network modeling to obtain responses for each required processability property (i.e., Mooney viscosity and number- and weight-average molecular weight); and (c) a simultaneous desirability response, due to several properties had to be optimized at a time. As a result, an 8% increase in productivity was accomplished, and some relationships were derived among CTA additions and polymer evolution, Mooney viscosity and molecular weights. The study was carried out to further apply this strategy in the production plant. The challenge was to keep the quality of the produced polymer, particularly in terms of processability. (C) 2011 Wiley Periodicals, Inc. J Appl Polym Sci, 2012