Abstract: Conventional inorganic hydrochemical parameters are difficult to distinguish the water source of coal seam roof water gushing in some mine fields. In order to solve this problem, a mine field in Yuheng Mining Area is taken as the research area. The organic indicators including TOC, UV₂₅₄ and dissolved organic matter（DOM）are added on the basis of inorganic indicators including K⁺+Na⁺, Ca²⁺, Mg²⁺, Cl⁻, SO₄²⁻, HCO₃⁻ and TDS to explore the function of inorganic-organic comprehensive index in identifying the water inrush source of coal seam roof. The main components and fluorescence intensity of DOM were obtained by fluorescence fingerprinting and PARAFAC. After principal component analysis（PCA）of the inorganic and inorganic-organic data sets of water samples, the inorganic index discriminant model and inorganic-organic comprehensive index discriminant model were constructed by using random forest algorithm（RF）. And the types of eight water samples to be tested are also correctly identified. The results show that the performance of comprehensive discriminant model is always better than that of inorganic discriminant model. Under the comprehensive model, the average precision, average precision, average recall and f1 harmonic index of RF algorithm reached 93.14%, 94.79%, 95.08% and 93.73% respectively, which increased by 9.71%, 11.84%, 12.25% and 11.5% compared with the inorganic model, and the back generation accuracy was 98.63%. The inorganic-organic comprehensive index can significantly improve the accuracy of identifying the water inrush source of coal seam roof.