Based on the investigation data of 124 sampling points in Quzhou county (the typical alluvial region of the North China Plain), spatial variation characteristics of soil particle composition were studied by the geostatistical component of ArcGIS (a kind of geographic information system tool). The results indicated that the trends of both sand and clay particle contents were second order in the east-west direction, and first order in the north-south direction. The trends of silt particle contents in the east-west direction and the north-south direction were second order and zero order, respectively. The anisotropy semivariance models of sand particle content showed that the ranges in the short axis direction were 17.83 km, 16.12 km and 21.16 km with zero order, the first order and second order trend effect, respectively. While the ranges in the long axis direction were approximate equal with all trend effects. The anisotropy semivariance models had longer ranges than the isotropy model with the first trend effects. It indicated that Kriging interpolation method considering anisotropy together with the first order trend effect was the best, by comparing the contour maps of soil sand particle content and the interpolation error of methods with different trend effects and anisotropy parameters.