Abstract: Based on the present status that the coal resources of many coal mine districts are becoming exhausted and the immense changes of aging mine water inflow in East China, this paper improves a new model(GM(1,1)-Markov-Information Renewal combination model) for forecasting mine water inflow and a comprehensive evaluation index(Z) for verifying the prediction by combination model. Verification results of models show that the combination model is superior to other models, because it reduces the error caused by the volatility of the serial data and the differences when replacing the old and new information, and can solve the impact of some uncertain factors(such as water-inrush accidentally and residual water burst, etc.) which affect prediction accuracy and dependability of aging coal mine water inflow over a long time span. The combination model and other models are used for forecasting mine water inflow of Jinggezhuang Mine of Kailuan Group, the results show that the comprehensive evaluation index of GM(1,1)-Markov-Information Renewal combination model is the highest(Z=0.475), and the predictive inflow values of mine water inflow in 2011-2015 will be 13.055 m³/min, 12.730 m³/min, 12.579 m³/min, 12.493 m³/min and 12.503 m³/min respectively.