Abstract：Because mine discharge sequence has both randomness and gray characteristics obviously,a GM( 1,2) prediction model was established to overcome shortcomings that the prediction effect of gray model was poor for long time series of large random fluctuations. Firstly,mine discharge sequences from July 2012 to December2013 of Wangxingzhuang mine were analyzed by using GM( 1,2) method,meanwhile,the aquifer level fall was considered when the GM( 1,2) model was established,and the fitting result was analyzed. The water inflow data from January to April 2014 were predicted and compared with the observed data,and GM( 1,2) was compared with GM( 1,1) on the forecast results. The results showed that the prediction accuracy of GM( 1,2) was97. 44%,and GM( 1,1) was 92. 60%. Prediction accuracy has been improved significantly using GM( 1,2).

Keyword：mine discharge;grey prediction;GM(1,2);aquifer level fall;