Abstract: Addressing the issues of delayed response, high false alarm rate, and reliance on manual labor for dust overrun disposal in the original ventilation system of Changdu Coal Industry, a "perception-analysis-decision-execution" closed-loop intelligent ventilation system for mines was constructed. The system adopts a "terminal-edge-cloud-application" architecture, with the "Ventilation and Prevention Data Cloud Manager" as the core to achieve multi-source data aggregation and governance. It integrates functions such as security situation analysis, dynamic calculation and optimization decision-making of ventilation networks, early warning linkage control, and dust monitoring linkage. Verification was carried out based on operational online data and typical fault simulation tests. The results showed that the average response time of the system's joint control was 0.59 seconds, the false alarm rate was reduced to 0.26%, the terminal online rate remained above 98.5%, and the data upload delay did not exceed 3 seconds. In terms of energy saving, the average load rate of the main fan was reduced from 86% to 72%, resulting in annual electricity savings of approximately 490 600 kW·h and annual electricity cost savings of approximately 345 000 yuan. The research results can provide a reference for the construction of similar intelligent ventilation systems in mines.