Abstract: Traditional research on disaster monitoring, early warning, and intelligent control in mines mostly focuses on the independent prevention and control of specific types of disasters, which fails to effectively address the nonlinear coupling risks of multiple disasters under complex conditions. To solve this problem, an intelligent ventilation system was used as the core carrier for mine disaster prevention and control. The intelligent collaborative control system of "One Ventilation and Three Prevention" in mines was analyzed, and a corresponding design scheme was proposed. This system adopted a high-precision ultrasonic full-section anemometer and a multi-agent reinforcement learning algorithm to achieve a wind speed measurement accuracy of ±0.1 m/s and a task-guided air volume adjustment with an error of ≤2%. Based on a dual-channel collaborative architecture integrating Spatio-Temporal Convolutional Neural Networks (ST-CNN) and Dynamic Bayesian Networks (DBN), the system fused multi-source data to construct gas safety zones, enabling early warning of gas anomalies and regional graded linkage. A real-time temperature field monitoring system was built through a LoRa wireless sensor network. Combined with disaster simulation and a three-level linkage mechanism, it reduced the disaster response time to the minute level. Relying on building information modeling and geographic information system technologies, a digital twin platform for the mine topological network was developed, realizing simulation and verification of the coupled evolution of the ventilation network and disaster prevention. A "data-driven decision-digital twin verification-equipment cluster collaborative control (3D)" technical path was established. Actual application results showed that the system could perform intelligent calculation of complex ventilation networks within 3 seconds, reduce the time required for disaster simulation and plan matching to 3-5 minutes, and achieve a delay of ≤500 ms in transmitting linkage control commands, effectively improving the intelligence level of disaster prevention and control and the emergency response capability.