Real time optimization(RTO) is widely used in the refinery industries. Fluid catalytic cracking(FCC) process is one of the most complex processes in petrochemical refinery. Due to complex reactions in FCC, the mathematical model of the FCC can't describe the plant perfectly. The RTO strategy based on mismatch model will lead to sub-optimal results. Modifieradaptation(MA) method is a new strategy solving the modelling mismatch in RTO. A basic MA method is applied to the FCC by using the forward finite-differencing gradient computation. The simulation results show satisfied performance of the MA method.