The understanding of the spatial variability of soil heavy metals is an important precondition for suitably monitoring and evaluating eco-environment quality in a primary agricultural production zone. 100 topsoils were sampled from the Zhengding County of the urban-rural transition zone in Taihang Piedmont Plain, China. The contents of eight heavy metals Cu, Zn, Cr, Ni, Pb, Cd, Hg and As were tested for each soil sample, and their spatial patterns were analyzed by using the semivariogram approach of geostatistics, with which the kriging method was used to estimate the unobserved points. Then GIS technology was employed to produce spatial distribution maps of the 8 elements. The results showed that the concentration of Cd exceeded its background level. The local pollution from Cd was attributed to the anthropogenic influence. The concentrations of the eight heavy metals are relatively lower than the critical values of the national soil quality standard. The correlation distance of soil heavy metals ranged from 3.28 to 11.63 km, with the eight heavy metals having moderate spatial dependence. Cu, Cr, Ni, Pb and As were associated with and controlled by parent material. The spherical model was fitted to the semivariograms of Cu, Cr, Cd, Hg, Pb and As, and the Zn and Ni were fitted with the Gaussian model and the linear model, respectively. The results are helpful for improving agricultural and forest ecosystem in the region.