Abstract: A snowmelt runoff model (SRM) was modified and applied to simulate the snowmelt runoff process in the Kaidu River basin in the Xinjiang Autonomous Region of China. The modified model was driven by the precipitation product of the Asian Precipitation-Highly Resolved Observational Data Integration Towards Evaluation (APHRODITE) of the Water Resources project. The daily effective active temperature was used in the model as the degree-days to replace the daily mean temperature for estimating the snowmelt rate. In order to reflect the effect of seasonal frozen soil on runoff, the accumulated active temperature was proposed to represent the state of seasonal frozen soil, and the runoff coefficient was modified using the accumulated active temperature. Using the meteorological and hydrological data observed from the Kaidu River basin, along with snow products from the Moderate Resolution Imaging Spectroradiometer (MODIS), the modified SRM was employed to simulate the runoff in the Kaidu River basin during snowmelt seasons. SRM was calibrated and validated using 2000 and 2006 data, respectively. Based on the coefficient of determination and the difference between simulated and observed runoff, calibrated and validated results simulated by the modified SRM using the daily effective active temperature as the degree-days were much better than that using the daily average temperature as the degree-days. The result shows that the interpolated APHRODITE precipitation data, the modified degree-days and runoff coefficient, expressed with the effective accumulated temperature as the SRM model inputs, can reflect the snowmelt runoff process in the Kaidu River Basin and greatly improve the snowmelt runoff simulation accuracy.