Based on the data about water consumption from 1986 to 2005, the water consumption of Xi′an city in 2010 and 2015 is predicted with the method of the MOBP model. This model unified the principal component analysis with the BP neural network. The results showed that, the driving forces of water resources was well analyzed and weighed by using the principal component analysis. The relative average error of the MOBP model in Matlab software was approximately 6％. The forecasting model and its corresponding calcualting method has higher forecasting precision and accuracy. The model predicted that the water consumption of Xi′an city in 2010 and 2015 was 24.14×108 m3 and 27.11×108 m3, respectively.

water resource；principal component analysis；MOBP model；Xi′an city