Abstract: The tailgate automatic control system is designed, and the water level data of the tailgate is collected in real-time by water level meter, setting the flap door in motion by the motor speed of feedback regulation for realizing water level control. Using the PID model as the motor speed controller, the water level control effect in the physical models for alluvial rivers and mountain rivers under the conditions of the fixed PID parameter is tested. The maximum deviation of the water levels in the mountain rivers reaches 1.3 mm, which can not meet the requirements of the specification. For this problem, the rule of adjusting PID parameters in controlling the tailgate by using the production learning method summarized, and a knowledge base with the water level of tracking error, level fluctuation velocity, expected fluctuation and tracking phase as the key elements is established with 12 control strategies. According to the measured data and the knowledge base rules, the corresponding control strategy is chosen independently to realize the by self-tuning of PID parameters when the tailgate is controlled. The practical application results show that the maximum deviation in controlling the water level by the tailgate control system based on PID parameters self-tuning is 0.5 mm.