Recent precision agriculture research has focused on the use of management zones (MZs) as a method for variable-rate fertiliser application. Topography factors have a profound influence on MZs. Delineating management zones by soil properties and elevations is therefore a practical approach. To achieve this objective, we selected a 3.1-ha typical mountain slope tobacco field in Sichuan Province in China, and collected 88 soil samples from the top 20 cm on an approximately 25-m grid of the study area, location and elevation were obtained by GPS. Geostatistics and the application of geographic information system (GIS) were necessary to combine soil properties and elevation for management zones delineation. Geostatistics analysis indicated that soil pH, OM, AK, TK, Fe and Zn were best represented through an exponential model; AN, AP, TP, Mn and Cu were spherical model. Principal component analysis (PCA) and fuzzy cluster algorithm were performed to delineate MZs; fuzzy performance index (FPI) and normalised classification entropy (NCE) were used to determine the optimum cluster number. Results showed that the optimum number of MZs for this study area was four and the elevation of each zone was relatively homogeneous. Analysis of variance indicated the heterogeneity of soil properties among the four MZs. We conclude that mountain tobacco fields can be managed in a site-specific manner by these methods.