Abstract: For accurate analysis of the spatial distribution and spatial variability of infiltration characteristics of aggregate gravel soil, the self-made double ring infiltration apparatus and gravel screening instrument have been used for testing gravel mass fraction of different particle size on slope , the permeability coefficient of gravel soil and different depth of gravel soil moisture content. Software is used for spatial interpolation of the discrete test data through the Kriging interpolation method. The cross tests are used to select a reasonable theoretical model, and the precise interpolation results are obtained. The results show that Kostiakov formula is used to simulate the infiltration process of gravel soil. The highest spatial interpolation precision about the gravel in different particle size sections is Gaussian semi-variable function theoretical model, and spherical model for the permeability coefficient of gravel soil and the moisture content of gravel soil at different depths. The gravels with particle size larger than 10 mm are mainly distributed in the top and middle slope regions, and the regional spatial variability of infiltration rate is relatively large. The small gravels (size of 2~10 mm) accumulate in the slope foot. The permeability coefficient of gravel soil along slope from top to bottom has a gradually decreasing spatial distribution trend. The moisture content at the slope waist is higher than that at the slope top and toe in shallow soil (0~15 cm), but that at the slope toe is higher than that at the slope top and toe in deep soil layer (15~25 cm). The water content of slope increases slowly with the increasing depth. The research results lay a theoretical foundation for the establishment of the prediction model for gravel soil.