Abstract: Conventional technology only considers one factor,which cannot achieve the same precision of gas outburst zone as multi-factor prediction methods. Taking an area as an example,the support vector machine ( SVM) network based on genetic algorithm was used to predict the gas content. The porosity was used as the discriminant factor of tec- tonic coal. Distribution of tectonic coal was obtained by probabilistic neural network (PNN). The quasi-density inver- sion method based on natural gamma curve was in-troduced to obtain the lithology of coal seam roof. Characteristics of gas content,tectonic coal distribution and coal seam roof lithology was comprehensively considered to establish the gas outburst risk area comprehensive fore-casting method,which provided a theoretical basis to determine the gas outburst danger zone. The prediction results were consistent with actual prominent positions,which proved that this comprehen- sive forecasting method had high accuracy in this study area.