Abstract: For problem that abnormal gas change below pre-set warning value cannot be found in time because existing safety monitoring and control systems of coal mine use overrun theory and take measured value by gas sensors as criterion, the paper proposed a pre-warning method for gas mutation below the pre-set limit value, and introduced two gas mutation pre-warning methods named static pre-warning and dynamic pre-warning. It did stratified sampling as well as distribution validation and fitting of static data and dynamic data of pre-warning basic data of gas mutation below the limit value according to probed gas data of N2202 working face of Yuwu Coal Mine, and got a conclusion that the basic data distribution of the static pre-warning and dynamic pre-warning is both in normal distribution. It also proposed a method to real-timely calculate upper limit value of the static pre-warning and dynamic pre-warning with hypothesis testing and interval estimation, which was verified by use of probed gas data of N2202 working face.