A conventional pan coating method was used to prepare propranolol-HCl sustained release coated pellets. Eudragit RS was used for controlling release. A 3(2) full factorial design was selected and nine experimental runs were performed to improve dissolution characteristics of the sustained release formulation. The independent variables were: plasticizer concentration in the coating and volume of coating dispersion applied to the pellets in the coating pan. Optimization was performed in order to maximize the fraction of propranolol released after 12 h with respect to constraints applied to the model after 1 h and 6 h release period.:The optimized formulation provided dissolution rates that were close to predicted values. A non-linear fitting of 'in vitro' dissolution data for the optimized formulation was performed. The kinetics of dissolution was shown to follow the Hixon-Crowell model. (C) 1997 Elsevier Science B.V.