The invention discloses an automatic high-risk area recognition method based on sub-graph isomorphism. The method includes the steps: (1) building an equivalent diagram model of a typical high-risk area, and completely expressing a topological structure and attribute information of the typical high-risk area; (2) converting a ventilation system to be analyzed and the typical high-risk area into an attribute graph form; (3) finding a portion isomorphic with the high-risk area in a ventilation network by the aid of a sub-graph isomorphism algorithm. The high-risk area in the coal mine ventilation network can be accurately recognized, a safety manager can be effectively assisted in analyzing the complicated ventilation system, and the management and control capacity of a coal mine for the ventilation system can be improved.