This work deals with the problem of dynamic optimization of an emulsion polymerization reaction in batch reactor. This process can be described as a hybrid system, i.e. dynamic system with both continuous and discrete character. We obtain optimal trajectories of control variable using control vector parameterization (CVP) method. This method translates infinite dynamic optimization problem into finite non–linear programming (NLP) problem. The NLP computation requires gradient information which we provide using method of adjoint variables.