Global warming induced by an increasing concentration of greenhouse gases in the atmosphere would change global and regional climate and take a significant influence on natural resources and environment. On the basis of water balance equation in unsaturated zone and digital elevation model, a monthly hydrological model char-acterizing heterogeneity of soil moisture content is developed. The model is calibrated and validated using monthly streamflow in the Yuanjiang basin in different spatial and temporal scales. Future possible runoff changes in the Yuanjiang basin is then simulated on the basis of the future climate scenarios predicted by global climate model (GCM). Simulation results demonstrate that annual runoffin the Yuanjinag basin is very sensitive to temperature rises and rainfall changes. From prediction results in the middle of 21th century presented by Hadley Centre for Climate prediction and Research HCCPR, UK (HADCM2), we know that mean annual temperature would increase 1.55℃ and precipitation decrease 0.43% in the Yuanjiang basin, and monthly rainfall would increase in flood season but decrease dramatically in drought season. Simulated results indicate that annual runoff would decrease 6.8 percent, and runoff in flood season increase 11% and in drought season decrease 47%, which results are not beneficial for flood defense and water resources utilization in the region.