Abstract: Based on the entropy weight method and fuzzy variable set theory,23 water samples were extracted from the four water sources at Shoushan mine and the thirteenth mine in Pingdingshan coalfield,a mathematical model of identification of mine water inrush was built,and the sources of 8 water samples were identified using the model. The results show that the probability of the 6 chemical indexes,namely Ka+ +Na+ ,Ca2+ ,Mg2+ ,Cl- ,SO24- and HCO- ,appea- ring near the Interval specific value is relatively high;the weights of Ka+ +Na+ and Ca2+ were 0． 270 8 and 0． 371 3 with entropy weight method,much higher than other index weight,and the sum of weights accounted for 64. 21% ,indi- cating the major role in the mine water inrush identification;the identification results of 7 groups of water samples were consistent with the actual type except the water sample 8,accuracy rate reaching 87. 5% ;for Pingdingshan coalfield, selecting the optimization criterion of parameter a = 2 in the identification model,the accuracy of entropy weight-variable fuzzy set model to identify the source of water inrush is higher;the model should be established on the basis of a- bundant spatial and temporal water quality data,and improved using the new information to ensure the identified model conforming to the reality and keeping pace with the times.