This paper compared the applicability of annual growth rate method, autoregressive model method and grey neural network method in the annual water consumption prediction of Beijing-Tianjin-Hebei region, and forecasted the annual water consumption of the Beijing-Tianjin-Hebei region from 2019 to 2025 based on the optimization method. The results showed that the annual water consumption of Beijing, Tianjin and Hebei Province showed different trends from 1997 to 2018. The annual water consumption of Beijing and Tianjin showed a non-linear trend of decrease and then increase, while the annual water consumption of Hebei Province showed a fluctuating decreasing trend. The grey neural network method is better than the other two models in Beijing, Tianjin and Hebei Province, so it is recommended as the first choice for the prediction of annual water consumption in this area. The prediction results of annual water consumption based on grey neural network method show that the annual water consumption of Beijing will be stable from 2019 to 2025, and the annual water consumption of Tianjin will increase slowly, while the annual water consumption of Hebei Province will continue to decline.