An understanding of spatial variability for soil chemical properties is essential in optimizing fertilization and sustaining yield. The objective of this study was to investigate the spatial variability of soil chemical properties in a 10-year-old jujube trees hillslope (2,725 m2) on the Loess Plateau of China. A total of 250 soil samples were collected at two depths (0-10 and 10-30 cm) in August 2010. Soil organic carbon (SOC), total nitrogen (TN), total phosphorus, pH, and electrical conductivity were determined for all soil samples, and data were analyzed by the classic and geostatistical methods. Classic statistics showed that soil organic carbon, TN, total phosphorus, and electrical conductivity had moderate variability, and pH had low variability. Soil organic carbon and TN in the lower position were significantly greater than in the upper and middle positions (P < 0.01), whereas pH in the middle slope was significantly greater than the other positions (P < 0.01). The geostatistical statistics results showed that the soil chemical properties were defined by the exponential or spherical models. The nugget-to-sill ratios indicated that all soil chemical properties had moderate dependences. The results of cross-validation showed that spatial prediction maps from ordinary kriging method were reliable in the study slope (the kriged reduced mean square error higher than 0.99 and the kriged average error close to 0). These results and maps could provide useful information for the development and application of precision agriculture on the Loess Plateau.