Abstract: In order to accurately predict the mine water inflow from Yuanbaoshan Surface Mine and to improve the safety coefficient of the coal mine production, based on the site measured data of several years’ rainfalls and coal productions in the mining area and the rainfall and coal production as the impact factors of the mine water inflow, a hydrogeologic analogy method and multi-element nonlinear regression analysis method was applied to individually establish the water inflow prediction model of the mining area. With the comparison analysis, a more accurate prediction method could be determined. The study results showed that the analogy value of the hydrogeologic analogy method would have a high difference to the site measured value of the water inflow, the correlation coefficient between the two values was -0.719 and was a negative correlation and the prediction accuracy was low. But the correlation coefficient between the predicted value of the water inflow obtained by the multi-element nonlinear regression analysis method and the site measured value was 0.946, the significance level test R2 was 0.894 and could explain the variation of the 89.4% water inflow. Thus the prediction would be more accurate and could be the basis to the prediction of the water inflow in the late mining area and to guide the mine safety production.