Abstract: According to characteristics of historical data and monitoring parameters in coal mine safety monitoring system, a rapid discovery method of information features of coal mine monitoring was proposed. The method uses compression algorithm of historical data based on error band to analyze sample data, discovers and stores information fragment that contains important features, analyzes meaning of the information fragment, conducts subject extraction and correlation analysis, researches correlation analysis of gas sequence, so as to get information features of important data in the coal mine safety monitoring system. The method has certain reference for completing scientific management of coal mine, excavating information of multiple sensors and law of gas discharge of coal mine.