This paper describes a statistical model-based control strategy for the regulation of the particlesize distribution in emulsion copolymerization. A database is generated using both a mechanisticmodel and experimental data. A partial least squares method is applied at each sample time togenerate the prediction model to be used in a predictive control algorithm. The proposed strategyis implemented in an experimental reactor in which the copolymerization of vinyl acetate/butylacrylate is carried out. The controller is effective in updating the input trajectories to regulatethe controlled distribution to a desired target in the presence of strong disturbances.