In order to investigate the temporal and spatial variation of soil moisture in the dry-hot valley, a study was carried out in the demonstration area of soil and water conservation at the old town village of Yuanmou, taking the Leucaena leucocephala plantation, Twig grassland and sloping land as the research object, by combining classical statistics and geostatistics analysis methods, the spatial and temporal variability of soil moisture on sloping lands of three land use types were studied. The results showed that: Soil water content in Yuanmou in the dry-hot valley was low (which was 7.56% and 12.80% in the dry and rainy season of forest land; 8.05% and 12.66% of grassland, and 19.37% and 22.95% in the farmland, respectively), soil moisture in the rainy season was significantly higher than that in the dry season, and followed the order of sloping farmland＞grassland＞forest land in both the dry and rainy seasons, and showed moderate variation (between 0.14 to 0.72); the autocorrelation coefficients of soil water content in different layers of soil of different land use types were all showed a positive to negative transformation trend, but the inflection point was different. The autocorrelation coefficients in the rainy season were higher than those in the dry season; the best fitting model for the forestland and grassland of soil moisture in the dry and rainy season were the same for different land use patterns (both the forest and grassland were spherical models during the rainy and dry season, and the farmland was the exponential model in the dry and rainy season), and all showed moderate or strong spatial correlation(between 0.05 to 0.39), which was greater in the dry season than in the rainy season. The spatial distribution patterns of soil moisture in different layers of soil of the same land type were similar in the dry season and rainy season, but the distribution patterns in the same layer of soil of different land use patterns were different.