Four Global Climate Model ( GCM) data ( GFDL,Had,IPSL and MIROC) recommended by the Inter-Sectoral Impact Model Inter-comparison Project drive SWIM,SWAT,HBV and VIC hydrological models to simulate runoff above the Cuntan hydrological station in the Yangtze River and to study the runoff changes under the global warming of 1.5℃ and 2.0℃. The results showed that: ( 1) In the 1.5℃ warming period,the hydrological model and GCMs simulated annual runoff increases of 5.5% ~ 8.3% and 3.5% ~ 11.4%, respectively; in the 2.0℃ warming period the hydrological model simulated a runoff increase of 4.8% to 6.7%. The IPSL simulated annual runoff showed a slight decrease,and the annual runoffs simulated by HAD and MIROC increased by 6.7% and 19%,respectively. The uncertainties from GCMs are 2.6 and 2.1 times that of hydrological models,respectively. ( 2) Under the conditions of two different global warming,the proportion of the monthly runoff collection averages is highly consistent with the proportion of monthly runoff during the reference period,but the proportion of the monthly maximum runoff under 1.5℃ and 2.0℃ global warming was 47.8% and 40.5% respectively,indicating that the monthly runoff contribution did not change significantly in the future warming period,but the extreme monthly runoff changed significantly. ( 3) Under 1.5℃ global warming, daily runoff increases in the dry and wet periods are 3% and 10%,but the change rates of runoff in the dry and wet periods are not large. Under 2.0℃ global warming,the increase in the dry season and the wet season was 3.6% and 8%,respectively,but the contribution rates of runoff in the dry and wet periods all showed a downward trend. Under 1.5℃ and 2.0℃ global warming,the flood flow once in 50 years ( P = 2%) will increase by 26.3% and 20.7% respectively over the base period. The floods in the 50-year period of the reference period will likely become once in 20 years,and the average annual maximum runoff volume for many years will also increase over the reference period.