Abstract: Based upon adaptive assessment of different GCMs recommended by IPCC, the future climate change scenarios were generated by using SDSM and ASD, respectively, and were used to drive the distributed hydrological model VIC and SWAT. The VIC was applied for simulating hydrological processes in the Taihu basin, which is selected as the typical watershed of the Yangtze River basin. The SWAT model was run for simulatimg hydrology in the upper reaches of the Yellow River basin. Then, the impact of climate change on hydrological cycle was quantitatively investigated. Results show that the methods adopted in this study for GCMs adaptive assessment and downscaling could reduce uncertainties effectively. It was detected that a decreasing trend in the upper reaches of the Yellow River basin; a slightly decreasing trend in the lower reaches, of the Yangtze River Basin,but with a significant increasing trend in the southeast of the Taihu basin during flood seasons for the future periods (2046-2065 and 2081-2100), comparing with the runoff in the baseline period (1961-1990). These results are of greatly significance for adapting climate change in different river basins for the future.