Abstract: In order to effectively prevent the disaster caused by coal spontaneous combustion in Dongrong No. 1 Coal Mine, the spontaneous combustion characteristics of coal seams in this coal mine were studied through the experiment of coal spontaneous combustion oxidation, the critical temperature of the marked gas in the experimental coal sample was determined and the variation of its volume fraction with temperature was analyzed; the principal component analysis method was used to comprehensively evaluate the nine indexes such as temperature, φ(CO), φ(C₂H₄)/φ(C₂H₆), etc., and the dominant index in predicting coal spontaneous combustion was selected. The results showed that the critical temperature of index gas and its regularity can reflect the spontaneous combustion process of coal. According to the principle of index gas optimization and the optimization results of principal component analysis, a prediction and forecast system for coal spontaneous combustion of Dongrong No.1 Coal Mine was established. φ(CO),φ(C₂H₆),φ(C₂H₄) and φ(C₂H₂) were selected as the main indicators and the ratio of φ(C₂H₄)/φ(C₂H₆) was selected as the auxiliary indicator. The accuracy of early prediction of coal spontaneous combustion was improved, and the prevention of mine fire was realized.