Abstract: Water inrush seriously restricts the mine safe production,and identifica water inrush sources quickly impor-tant for the prevention and control of mine water inrush. This research was carried out to build a mine water inrush i-dentification model based on dynamic weight-set pair. Six parameters including K+ +Na+ ,Ca2+ ,Mg2+ ,Cl- ,SO24- and HCO- were elected as recognition factors and source identification ranges were determined according to 53 groundwater samples extracted from five water sources at Hulusu coalfield. Then,10 mine water samples with definite source to test dynamic weight-set pair model. At last,this established model was used to predict mine water inrush of 21102 working face of Hulusu coalfield on April 26,2015. The results indicate that dynamic weight method is reasonable because sub-jective weights and objective weights are taken into account,simultaneously. On one hand,the influence of human fac-tors is weakened,and on the other hand the actual situation of the identification parameters is considered. The weights of SO24- ,Cl- ,and K+ +Na+ are 0. 38,0. 25 and 0. 20 using dynamic weight method much higher than other parameters weights,and the sum of weights account for 83% of the total,indicating their conclusive roles in the mine water inrush identification. The verification results of 9 groups of mine water samples are entirely consistent with the actual type. Only one Quaternary water sample identification results inconsistent with the reality,this is because there are few Qua-ternary samples in modeling and the samples to be tested exceed the recognition interval. the prediction result of 21102 working face of Hulusu coalfield is consistent with actual mine water inrush result. This is because almost all water in-rushes of 21102 working face comes from Zhiluo and Cretaceous aquifer. large number of Zhiluo and Cretaceous groundwater samples are used to build model to establish appropriate identification intervals. Dynamic weight-set pair model should be established on the basis of abundant water quality data to ensure accurate identification ranges and valid water inrush source.