The control in over tolerance length in strip head and tail segment in cold rolling is a difficult problem urgently needed to be solved in cold rolling strip production. The head and tail segment of cold rolling was of lager parameter perturbation and easily influenced by the external factors, and it was difficult to establish the precise mathematical model, which resulted in over tolerance thickness of head and tail segment, low rate of finished products. A control approach based on fuzzy rule switching was proposed, which combined the fuzzy adaptive and traditional PID control so as to improve the thickness gauge control accuracy of the head and tail segment. Simulation and experimental results show that the approach improves the thickness gauge control accuracy of the head and tail segment, decreases the over tolerance length and avoids the system instability caused by the routine threshold switching, and raises the rate of finished products.