Abstract: In order to realize the real-time monitoring, prediction and early warning of coal and gas outburst danger in coal roadway heading face, Malan Mine has installed and deployed the KJ1521 coal and gas outburst prediction and early warning system, and we carried out the research on the difference distribution of time-frequency domain parameters of micro-seismic signals under different scenarios and outburst danger response characteristics. The results show that under normal monitoring, there are differences in the quantitative distribution of time-frequency domain parameters of micro-seismic signals in different regions, but the qualitative laws are basically the same. The time-frequency domain parameter distributions of micro-seismic signals in special scenarios such as micro-seismic sensor movement, non-vertical installation of sensor and sensor short circuit correspond to low and high value anomalies of waveform amplitude and duration, low value anomalies of rise time, and low value anomalies of peak frequency 50 Hz and below, respectively; the reasonable threshold range set for the time-frequency domain parameters of the micro-seismic signal can effectively filter out the noise signal caused by the abnormal state of the sensor, and improve the accuracy of the response of the micro-seismic signal to coal and gas outburst. When the coal roadway heading face encounters fault structure, the rise time mean and duration mean of effective micro-seismic signal show high value anomaly, the amplitude mean shows local high value fluctuation, and the peak frequency mean shows low value anomaly. The average amplitude, rise time, duration and peak frequency of micro-seismic signals show the precursory characteristics of high value anomaly when the roadway is excavated through stress concentration area.