Based on the daily meteorological data of 50 weather stations and daily runoff data from the Cuntan Gauging Station, HBV and SWAT hydrological models were calibrated and validated by using DEM, land use data and soil texture information. Possible impact of future climate change on river runoff was analyzed through forcing HBV and SWAT models by CCLM dynamic downscaled climatic data. The results show that both of HBV and SWAT models are suitable for simulating monthly river runoff in the Cuntan catchment, with NashSutcliffe efficiency coefficient of about 0.90 in the calibration and validation period. But it is also clear that the SWAT is good at simulating floodpeak value while HBV has good performance on baseflow value. Compared with baseline period (1986-2005), mean temperature, maximum temperature and minimum temperature in the Cuntan catchment will increase obviously with persistent rising trend by CCLM in 2011-2040 under RCP4.5 scenario. Annual precipitation is also expected to rise a bit in 2011-2040, but will decrease slightly since 2030. From the ensemble mean of the two hydrological modeling results and changes in runoff quantile, we can see that the annual runoff will increase 14.2% with probability of higher flood volumes in 2011-2040 than 1986-2005.