Abstract: A detection and early warning trigger mode of coal mine underground safety event based on event driven was constructed by using complex event processing technology. By adopting underground RFID personnel positioning data and environmental monitoring data as application data, a largescale underground safety flow data processing framework based on complex event processing technology was built. And then a double layer parallel mechanism for self learning of offline data association rule and self matching of online data matching tree structure was designed, so as to achieve high efficiency detection and early warning trigger for early warning of underground safety event. The test results show that the detection and early warning trigger mode of coal mine underground safety event based on event driven has significant advantages in abnormal event detection and detection efficiency compared with safety monitoring model based on relational database.