Abstract: In view of problems that existing geophysical prospecting and drilling methods of coal mine are not effective in advanced detecting of geological anomalies such as small geological structures and coal seam occurrence changes, as well as insufficient exploration and utilization of hidden information in outburst prevention and prediction data, the idea of geological anomaly intelligent identification according to association between outburst prediction characteristics and geological anomalies is put forward. From data distribution of a single coal and gas outburst prediction event and the data change of several consecutive coal and gas outburst prediction events, 10 indexes of outburst prediction characteristic are constructed, forming the coal and gas outburst prediction characteristic index system. Applying the method of association analysis, a method of geological anomaly intelligent identification based on coal and gas outburst prediction characteristics is proposed, and the key technologies are emphasized, such as binary attribute transformation of characteristic index, association rule analysis, effective rule extraction, identification criteria establishment and geological anomaly possibility levels classification. A geological anomaly intelligent identification system based on coal and gas outburst prediction characteristics is designed and devoloped using B/S structure, which achieves online collection of outburst prediction information, automatic analysis of outburst prediction characteristics, advanced dynamic identification of geological anomaly, and joint distribution of identification result via multi-channel of website and mobile terminal. The field test results show that the system can independently construct the geological anomaly identification criteria, and the total accuracy rate of geological anomaly identification reaches 87.63%, which provides an effective means for the coal mine to grasp the geological anomaly in advance, and realizes extended application of hidden value of the outburst prediction data.