Aiming at improving the performance of clamping force control of metal V-belt Continuously Variable Transmission (CVT), such as metal belt abrasion and low efficiency, a clamping force fuzzy-PID controller model was established. Considering the defect of fuzzy-PID controller, such as subjective, rough and imperfect, genetic algorithm was applied to off-line optimize both membership function and fuzzy rules. At last, road tests were done to verify the performance of the optimized fuzzy-PID controller. The test results have shown that the improved fuzzy-PID controller has a satisfactory effect on the clamping force control with better tracing precision and less steady state error. Additionally, the CVT's fuel economy was enhanced by reducing the safety factor of clamping force from 1.3 down to 1.24 on the basis of control precision.