Abstract: The environment of Shule River basin is sensitive to climate change. Therefore, this study tries to predict the climate change over this basin using the SDSM model （a statistical downscaling model） based on the observation data of 6 meteorological stations in (or adjacent to) the basin and the output of HadCM3. The research results show that the Nash coefficient is all above 0.98 for the air temperature prediction; while the best precipitation prediction occurs at the Tuole meteorological station, with the Nash coefficient of 0.6, and the predicted precipitation is overall higher than the observed precipitation. The comparisons of the inter-annual values between the observed and predicted precipitation show that the SDSM model has a good prediction capacity for the tendancy of the air temperature, but not for the precipitation. The prediction shows that, compared with 1961—2001, the period 2020—2039 has a higher air temperature by （0.8~1.9） ℃ under A2 scenarios and （1~2）℃ under B2; and there is a similar change in the precipitation under both A2 and B2, namely 54 mm decrease at the Tuole meteorological station and 6 mm increase at the Mazongshan meteorological station. Remarkably, the studies indicate that in the Shule River basin, the prediction factors which are interrelated highly with the prediction factors are located not on the grid of the meteorological stations (except the Tuole meteorological station), but on the eastern one, and the reason is to be further studied.