The generalized delta rule (GDR) algorithm with generalized predictive control (GPC) was used to control the temperature of a jacketed batch reactor in which styrene polymerization occurs under isothermal conditions. The effects of different optimal conditions were examined on monomer conversion, average viscosity molecular weight and chain length. The neural network model based on the relation between the reactor temperature and heat input to the reactor was used. The efficiency of the GDR with GPC was examined by simulation and experimentally using GDR parameters specified at constant temperatures, and compared with Self-Tuning PID (STPID). It was observed that the control experiments provided a good performance in maintaining the reactor temperature at its set point and yielded polymer product with desired properties.