Abstract: In view of the defects of the uncertainty measure comprehensive evaluation mathematical model in the determination of weight value and the identification of confidence degree, on the basis of the analysis and research on many random and fuzzy uncertainties in the process of evaluation of the groundwater pollution risk for abandoned mines, the mathematical model for comprehensive evaluation of uncertainty measures is improved by using the combination weighting of game theory and the confidence recognition of distance discrimination, an improved uncertainty measure theory based on the combination weighting of game theory is proposed to evaluate the groundwater pollution risk in abandoned mines. In this model, the linear uncertainty measure function of each evaluation index in the evaluation of the groundwater pollution risk for abandoned mines is constructed based on uncertainty measure, the subjective weight based on analytic hierarchy process and the objective weight based on entropy method are optimized by game theory, and the optimal weight of index is determined. Optimization credible degree recognition criterion by distance discrimination is used to evaluate the groundwater pollution risk of abandoned mines. This model is applied for the evaluation of groundwater pollution risk in Hongshan-Zhaili abandoned coal mine. The results show that evaluation results reflect the comprehensive situation of underground pollution for abandoned mine, and has good practicability. It provides a new idea for the evaluation of groundwater pollution risk in abandoned mine.