Abstract: In order to study different industrial analysis indexes affected to the length of the explosion flame,with the application of different characteristic coal samples,a full automatic industrial analysis component test device and the coal dust explosion flame length test device were applied to measure the moisture content,ash content,volatile matter and explosion flame length of the coal samples. The paper had an analysis on each industrial analysis index affected to the length of the explosion flame. With the assistance of the support vector machine,the test data were applied to establish a classified prediction model to predict the grade of the explosion danger. The results showed that when the coal dust particle size was within a certain range,the volatile matter content increased and the moisture and ash contents decreased both would promote the length increased of the coal dust explosion flame. When the dust particle size was in a certain large degree,the length of the coal dust explosion flame would not be affected by the industrial index. The predicted result accuracy of the classified prediction model was 95.83%. The method could be applied to predict the explosion danger grade and could be assistance to the evaluation of the coal dust explosion risk.