Abstract: At present, artificial intelligence technology has achieved certain application results in various fields such as underground comprehensive mining, excavation, electromechanical, transportation, and ventilation in the coal industry. However, there is a lack of system level solutions based on artificial intelligence technology, and most of them remain at the level of single scenarios and business applications. In order to improve the level of coal mine safety production, it is urgent to achieve the integration and data sharing of various systems for excavation, electromechanical, transportation, and ventilation in the coal industry, to meet the data extraction and application needs of artificial intelligence scenarios. By building AI nodes in the cloud, edge, device, a collaborative architecture system based on cloud, edge, device established, and an autonomous and controllable artificial intelligence platform based on industrial ring network, 5G underground in coal mine, industrial control data, big data, private cloud, robots, and intelligent perception agents is constructed. The initial formation includes infrastructure, AI development framework, dataset, AI training, AI deployment The AI service capability and business application of the Shendong mining area's artificial intelligence platform architecture are bottom-up. Supervised learning, semi supervised learning, transfer learning and other technologies are applied to improve the efficiency and quality of model training. Training data and validation datasets are collected from monitoring points deployed at some production sites in Shendong mines as research objects. The relevant trained AI models and algorithms are deployed to the scene of coal mine safety production, further improving the intelligence level of coal mine expert systems, robots, decision management, safety management and equipment monitoring. Taking the example of pedestrian-vehicle non-concurrent movement, industrial camera ambiguity, and main transportation system safety monitoring in Shendong mining area, the application effect of the artificial intelligence safety production control platform is verified.