[Objective] Soil moisture characteristics under four vegetation restoration treatments were investigated at different locations of the Hangjinqi photovoltaic power station in the Kubuqi Desert， Inner Mongolia， China， to provide a scientific basis for the rational configuration and management of sand-fixing vegetation in areas with photovoltaic power stations. [Methods] The cutting-ring method was used to determine soil water content at 0—100 cm depth under four vegetation types between and under the panels at different locations （windward side， array belly and leeward side） in the study area and to analyze soil moisture vertical distribution， soil moisture variability， soil water storage and soil water storage deficit. [Results] ① All four vegetation restoration measures increased soil water content， with the average water content in the order of Artemisia ordosica （2.09%） > Leymus chinensis （2.08%） > Glycyrrhiza uralensis （2.06%） > Hedysarum scoparium （1.98%） > bare sand （1.34%）. Spatial distribution was in the order of windward side < leeward side < array belly and interpanel < subplate. Soil moisture was higher under G. uralensis and L. chinensis vegetation restoration than under A. ordosica and H. scoparium vegetation restoration. ② The range of change in soil water storage at 0—100 cm depth under different vegetation restoration measures was 16.72—51.55 mm， and the soil water storage trend was consistent with the soil water content trend. ③ Soil water storage deficit under the different vegetation restoration measures in the photovoltaic power station was in the following order： bare sand （78.81%） > H. scoparium （69.00%） > L. chinensis （68.95%） > A. ordosica （67.23%） > G. uralensis （63.11%）， and decreased with an increase in soil depth， and the spatial distribution was in the order of inter-slab > sub-slab， windward side > leeward side > array belly. [Conclusion] The vegetation restoration measures in the present study increased soil water content in all parts of the soil， and A. ordosica and H. scopariumk had greater effects than the other measures in the 0—30 cm depth range， with lower coefficients of variation. Soil water content， water storage， and water storage deficit under L. chinensis and G. uralensis were superior to those under the other species treatments at depths < 30 cm. Therefore， during the operation and maintenance period， planting A. ordosica and H. scoparium on the windward side， with severe wind erosion， could effectively alleviate wind erosion， and L. chinensis， G. uralensis and other cash crops should be planted in the hinterland and other areas of the array that are less severely affected by wind erosion to address the problem of severe secondary extinction and achieve sustainable development.