Abstract: The paper detailedly analyzed problems of how to determine lag time of data between two sensors, effect of sensor error, accuracy rate of alarm in application of alarming method for abnormal data of coal mine gas based on three kinds of correlation models of difference between the maximum and the minimum, the minimum and correlation coefficient, and proposed that data curve comparison method should be used instead of distance anemometer algorithm to determine the lag time, meanwhile, found phenomenon that movement speed of gas air mass in roadway is lower than wind speed and gave preliminary explanation for the phenomenon. It proposed a method to use mathematical statistics method to eliminate effect of sensor error. Combined with field data, verification was conducted on accuracy rate of the three kinds of correlation models, and the result shows that the alarming method can be effectively applied in practice.