运用传统的和改进光滑度后的GM(1,1)模型分别对南京市年用水量进行建模预测。改进的GM(1,1)模型先对原始数据进行线性变换,再对变换后的数据应用GM(1,1)模型进行预测,有效提高了模型的拟合精度和预测精度。1998～2006年南京市用水量的建模结果分析表明,改进后的GM(1,1)模型预测用水量时拟合效果更好;2007～2009年用水量资料验证亦表明改进后的模型预测精度更高。 The traditional model and GM (1,1) model with improved smoothness are respectively used to predict the annual water consumption in Nanjing. The improved GM (1,1) model first linearly transforms the original data and then applies the GM (1,1) model to the transformed data to effectively improve the fitting accuracy and prediction accuracy of the model. The results of modeling of Nanjing water consumption from 1998 to 2006 show that the improved GM (1,1) model fits well when predicting water consumption. The water quality data from 2007 to 2009 also show that the improved model predicts Higher accuracy.