Abstract: Deep mining fault areas are the key points of coal mine management of safe production. At present, in the process of coal mine site safety production management, the safety risk assessment before roadway crossing fault mostly comes from experience judgment, qualitative estimation. Based on the analytic hierarchy process(AHP), multiple factors and their weights affecting the safety risk of roadway crossing fault are scientifically determined, and a comprehensive ranking of the safety risk status of 16 heading surfaces has been made. The grey clustering method is applied to establish the risk assessment and prediction model of crossing fault safety, and the safety risk grade of each heading surface is evaluated. The results show that it has good consistency to evaluate the safety risk of tunneling roadway crossing fault by AHP comprehensive ranking and grey clustering method, the evaluation process is based on the geological data obtained from the mining process, and the data processing is relatively simple, which has a positive reference significance for the study and solution of similar problems in coal mines.