In order to accurately identify the Smith predictor model in the Smith-fuzzy PID temperature control system of the rubber extruder,and to better realize the hysteresis compensation control of the temperature control system,the Smith-fuzzy PID temperature control system was built with MATLAB software,In order to avoid the phenomenon of “data saturation”,the FFRLS identification algorithm was used to identify the new and old data of the Smith predictor with different identification reliability,and the control effects of the system were compared before and after the FFRLS identification optimization.The experimental results show that the Smith predictive controller after FFRLS identification has better delay correction ability and parameter change identification ability under the effect of different interference;the Smith-fuzzy PID temperature control system of the rubber extruder using FFRLS identification is more robust,which can improve the control accuracy of the temperature control system.