Abstract: Cold areas are primary runoff generation areas of China's major rivers.The accuracy of their runoff forecasts directly affects the scheduling of downstream water conservancy projects and the water resources management.This study aims at the problem that traditional conceptual hydrological models rarely consider the combined effects of snowmelt and soil freezing and thawing processes on the runoff process.Based on the Xin'anjiang model and using precipitation and air temperature as driving data, this study proposes a conceptual soil freeze-thaw module based on the equivalent soil temperature and soil freeze-thaw characteristic curve to describe the runoff process under the joint influence of snowmelt and soil freeze-thaw, and a Xin'anjiang model considering snowmelt runoff and soil freeze-thaw is developed.Taking the watershed above the Ganzi hydrological station in the Upper Yalongjiang River basin as a typical research area to carry out the daily streamflow simulation during the snowmelt period of 2017—2022.The results show that, compared with the HBV model, the Xin'anjiang model with snowmelt runoff and soil freeze-thaw processes can significantly improve the accuracy of streamflow simulation of the Ganzi station during the snowmelt period.The median Nash efficiency coefficient increases from 0.69 to 0.83, and the median correlation coefficient increases from 0.88 to 0.92.The Xin'anjiang model improved in this study that considers snowmelt and soil freezing and thawing can provide support for hydrological forecasting in cold regions of China.