Abstract: A total of 163 soil samples (0-20 cm layer) were collected from the grid sampling plots (80 m×80 m) in Huanjiang Observation and Research Station of Karst Ecosystem in a small catchment in Karst clusterpeak depression area, South China. By using classical statistics and geostatistics, the spatial heterogeneity of mineral components (SiO₂, Fe₂O₃, CaO, MgO, Al₂O₃, MnO, and TiO₂) in the soils were studied. The contents of the  seven soil mineral components  in the study area differed greatly, being in the order of SiO₂>Al₂O₃>CaO>MgO>Fe₂O₃>TiO₂>MnO, and the variance coefficients also varied obviously, in the order of CaO>MgO>Fe₂O₃>TiO₂>SiO₂>Al₂O₃>MnO. The seven mineral components accounted for 69.4% of the total soil mass. The spatial patterns and the fittest models of the seven soil mineral components differed from each other. All the seven soil mineral components had a strong spatial autocorrelation, with shorter variation ranges and stronger spatial dependence. The Kriging contour maps indicated that the distribution patterns of soil SiO₂, Fe₂O₃, Al₂O₃, MnO, and TiO₂ were similar, being higher in south and east, lower in north and west, higher in depression, and lower in slope, while the distribution patterns of soil CaO and MgO were in adverse. Natural conditions (vegetation, bare rock rate, slope degree, and slope aspect, etc.) and human disturbance were the most important factors affecting the spatial patterns of the soil mineral components.