Aiming at the problems of high energy consumption, lagging regulation and prominent safety hazards in the traditional mine constant frequency constant air volume ventilation mode, this paper starts from the aerodynamic characteristics of the ventilation network, analyzes the energy dissipation mechanism of the traditional regulation mode and the energy saving principle of frequency conversion regulation, and constructs a closed-loop control system of ' global perception-feedforward prediction-intelligent decision-frequency conversion execution '. The error back propagation neural network is used to establish the air volume demand prediction model, and the fuzzy adaptive PID control strategy is combined to realize the dynamic adaptive adjustment of the main fan speed. Taking Huayang No.1 Mine as the engineering practice object, the hardware reconstruction and intelligent algorithm deployment of the ventilation system were completed, and the one-month actual measurement verification was carried out. The results show that the system can control the gas concentration of the working face stably at 0.4% ~ 0.5%, and completely eliminate the hidden danger of gas overrun. The average daily power consumption of the main fan is reduced from 34 200 kW·h to 24 820 kW·h, and the comprehensive power saving rate is 27.4%. The research provides theoretical support and engineering paradigm for intelligent transformation of mine ventilation and energy saving and consumption reduction.