Abstract: Based on the mechanism of coal and gas outburst and the physical characters of outburst process, in combination with a coal mine safety monitoring system, early warning indexes were proposed by the time series of methane concentration and coal temperature. An early warning model of coal and gas outburst was built by globe information system (GIS) technology and artificial neural network. Taking 17180 working face for an example, the early warning model was tested in an exact mine. The result shows that the model can provide technical support for early warning of coal and gas outburst.