Once the conventional PID control at the object changes, the parameters of the controller cannot be automatically adapted, and it cannot well meet the requirements of thickness control precision of cold rolling mill. According to the problem, a fuzzy adaption PID controller is designed, the fuzzy inference method is used to achieve the PID parameters online self-adaptive to realize AGC stability and the precision of gauge control system is enhance. The results of simulation indicate that fuzzy PID has a better adaptive capacity, faster response, and stronger robustness. It gives a good control performance to the thickness control.