Abstract: Spatial variability of soil nutrients provides useful information for improving agricultural practices and ecological management. The objectives of this study were to (i) quantify the spatial variability and spatial correlation of soil particle composition across areas and (ii) generate contour maps for the soil particle composition to reveal the characteristic of distribution. Grid sampling method (100 m×100 m) and geostatistics were applied to analyze the spatial distribution of soil particle composition in tobacco plantation field in Pingdingshan. The results showed that, sample variograms of soil particle composition were fitted well by spherical and exponential models, respectively. Sand and clay had a strong spatial correlation, good structural, while the silt with a moderate spatial correlation and bad structural. The anisotropic of three components of soil particle composition were little. The clay had the strongest of trend effect and the biggest of spatial variability, sand had little spatial variability with certain trends in effects, silt had a little trends effect. Kriging interpolation results showed that the high sand content in the north, while the lower in south-west; the distribution of clay and silt content was converse to sand. In the study area, the most account of the area was loam, the clay loam had small distributed only in the southwest corner.