Soil moisture is the core component that links the earth's surface water cycle， energy cycle and biogeochemical cycle， and it is also the key parameter to study the ecosystem in karst rocky desertification area. The aim of this study was to retrieve the soil moisture with multi-temporal Sentinel-1 C-band SAR data and observation equations were constructed by using the Alpha approximation model over karst rocky desertification area. The spatial and temporal variation characteristics and the error influencing factors of the soil moisture retrieval results were analyzed. It is found that the overall change trend of soil moisture in the observation period is highly consistent with the trend of rainfall change. The maximum value of soil moisture and the degree of spatial heterogeneity in karst rocky desertification area is significantly higher than those in non-rocky desertification area. The results were validated using ground measurements of one acquisition date， with root mean squared error （RMSE） value of 0.059 cm³/cm³ and mean bias value of 0.026 cm³/cm³. The method of retrieval of soil moisture by Alpha approximate model is applicable to karst rocky desertification area， the mixed pixel problem caused by the surrounding complex habitat conditions of the surface soil is the main influencing factor of inversion error. These results can provide reference to obtaining soil moisture under mountain complex environment by multi-temporal observation method， and provide support for ecosystem restoration and ecological industry development in karst rocky desertification area.