Abstract: In order to estimate the water resources of a river basin under changing conditions by simulating the hydrologic station monthly runoff, a hydrology model was established based on the wavelet neural network using observed meteorological factors to simulate runoff process in the upper Hanjiang River, and according to the future climate change incremental scenarios, runoff response process at the Shiquan hydrologic station was analyzed at different time scales. The wavelet neural network model by automatic learning and training can be used to simulate the reliable accuracy runoff data obtained from the Shiquan hydrologic station at the upper Hanjiang catchment based on the monthly precipitation and average monthly temperature. The simulated results show that, based on the model and different climate change scenarios, the increase in the annual average runoff is significant under the different scenarios, the maximum range of the annual average runoff is from -34.7% to 21.4%. In the case of no changes in rainfall and the rise in temperature, the mean annual runoff variation ranges are from -5.1% to -13.3%. The rise in temperature caused significant increase in the winter runoff, and the spring and autumn runoff also have the decreasing trends, and it is more significant in the autumn, but the rainfall changes have a significant influence on the summer runoff.