A Real Time Optimization (RTO) algorithm is proposed to optimize the reactors' propane content for the Spheripol Polypropylene process. This algorithm improves the present approach of arbitrarily selecting "optimum" propane values, based on monthly/yearly average balances. The optimum propane content depends on factors such as the costs of offgas and catalysts, the propane content in the feedstock and the catalyst mileage. The steady state process model is developed and tested over plant data to prove its predictive capabilities. Results from the off-line, as well as the in-line implementation of the algorithm are presented that demonstrate its applicability. Estimates on the amount of the expected savings for a medium size Polypropylene plant are also presented. (C) 2021 Elsevier Ltd. All rights reserved.