Abstract: Reasonable micro-seismic early warning indicators and early warning model are the basis of coal and gas outburst prediction by using micro-seismic monitoring technology. By comparing the seismic monitoring data measured in field, the results show that the energy rate of micro-seismic can reflect reality better than event rate. The energy rate can be used as the main index, and the event rate can be used as the auxiliary index, both of which comprehensively evaluate the outburst danger in front of the driving face. Based on the approximate normal distribution of micro-seismic warning indicators and the Russian micro-seismic warning criterion, the comprehensive early warning model of micro-seismic prediction outburst is proposed, the two criteria of which contain two aspects and complement each other. Through a large number of field data, the accuracy of the comprehensive early warning model is verified. In addition, micro-seismic monitoring data before and after the outburst shows that the maximum value of the early warning index appears while the outburst occurs, and there is the quiet period of about 10 days before the accident.