Fractal dimension is a parameter that quantitatively describes the complexity of fractal set. In order to understand the spatial variation characteristics of soil fractal dimension and its influencing factors, different geomorphic units in a typical basin of the North China plain were selected for sampling. And the grain size distribution of 180 soil samples were measured by laser diffraction. Fractal dimension was analyzed by the software of SPSS26 and Voxler4.3 to search for significant influencing factors. Then, in the same coordinate system as fractal dimension, models of soil particle sorting were built. There was a decreasing trend from the sloping plain at the southwest piedmont (1.6244) towards the alluvial plain (1.2164) about fractal dimension. In piedmont plain, fractal dimension was significantly correlated with the minimum particle size and proportion of clayey particles, respectively. In coastal plain, fractal dimension was extremely significantly correlated with the maximum particle size and the proportion of sandy particles, respectively. The spatial variability of soil structures in the study area was significant. The models features of sorting and fractal dimension are similar, and a linear relationship existed between fractal dimension and standard deviation as follows: D = 1.236σ₁ + 0.082. This research supports quantitative and visual characterization of spatial variability of soil structure and sorting of soil particles.

分形维数是一个定量描述分形集复杂度的参数。为了了解土壤分形维数的空间变化特征及其影响因素，选择了华北平原典型盆地不同地貌单元进行采样。并通过激光衍射测量了180个土壤样品的粒度分布。分形维数通过SPSS26和Voxler4.3软件进行分析，以寻找重要的影响因素。然后，在与分形维数相同的坐标系下，建立了土壤颗粒分类模型。分形维数从西南山前的倾斜平原（1.6244）向冲积平原（1.2164）呈下降趋势。在皮埃蒙特平原，分形维数分别与最小粒径和黏土颗粒比例显着相关。在沿海平原，分形维数分别与最大粒径和沙粒比例极为相关。研究区土壤结构的空间变异性很显着。分形维数和分形维数的模型特征相似，分形维数与标准差之间存在线性关系，其关系如下：D =1.236σ₁  + 0.082。这项研究支持对土壤结构空间变异性和土壤颗粒分类进行定量和视觉表征。