Abstract: The coal spontaneous combustion disaster seriously threatens the mine safety production. Mastering the correlation between the micro group and characteristic temperature of coal spontaneous combustion plays a positive role in the prevention and control of coal spontaneous combustion disaster. In order to determine the change and correlation of characteristic temperature and coal microstructure in the process of spontaneous combustion of coal with different metamorphic degrees, the mass change characteristics and microstructure of four coal samples with different metamorphic degrees were tested by thermogravimetric experiment and Fourier infrared spectrum experiment, and the relationship between them was determined by Pearson correlation coefficient method. The results show that with the increase of metamorphic degree, the content of aromatic hydrocarbons and aliphatic hydrocarbons in coal increases and the content of oxygen-containing functional groups decreases. The groups with the highest content in the three types of groups are C=C, —CH₃ and —OH respectively. The weight loss curve and characteristic temperature of high metamorphic coal gradually move to the high temperature zone, and the spacing between characteristic temperatures increases. The characteristic temperature of coal spontaneous combustion increases with the increase of aromatic hydrocarbon content and decreases with the increase of oxygen-containing functional group content. The higher the content of aromatic hydrocarbons, the more difficult coal spontaneous combustion is, while the oxygen-containing functional groups are just the opposite. The calculation results of Pearson correlation coefficient show that with the increase of temperature, the functional groups with the largest positive correlation coefficient are C—H, C—H, C—H, substituted benzene, C=C and C=C, and the functional groups with the largest negative correlation are COO—, COO—, C—O—, COO—, oh and —Oh, respectively. On the whole, C—H has the greatest positive correlation with the characteristic temperature, while COO— has the greatest negative correlation with the characteristic temperature. The change of characteristic temperature has an obvious linear relationship with the content of the maximum functional group. Based on the results of Pearson correlation coefficient, a multiple linear regression model of coal spontaneous combustion tendency is established by using the contents of C—H and coo functional groups which have the greatest correlation with characteristic temperature.