The patterns and controls of soil organic carbon(SOC) and total nitrogen are critical for our understanding of recycling of nutrients. Spatial heterogeneity causes uneven soil resource distribution. Quantification of the spatial variability is essential for evaluating attributes at unsampled locations. This study was conducted in Luya Mountain, Shanxi Province, China. Four sample plots (30m×30m) were placed, two subalpine meadow plots (Plot A, 2756.3 m; Plot B, 2542.3 m) and two Picea meyeri forest plots (Plot C, 2656.8 m;Plot D, 2387.2 m), The spatial heterogeneity of soil organic carbon(SOC) and total nitrogen(TN) was analyzed based on theory and methodology of spatial pattern analysis in geostatistics.. Soil samples (n=119) were collected from each of the plots in the summer of 2010. We calculated the isotropic semivariograms of SOC and TN. Then spherical models were used to test the semi-variances of SOC and TN for spatial dependence. The experiment results indicated that the SOC of four plots were 49.84 g/kg (A), 38.33 g/kg (B), 47.06 g/kg (C), and 40.67 g/kg (D), respectively. The SOC contents appeared to be higher in the plots with higher elevation. Rather, the spatially variance showed more intensively in the plots with lower altitude. In contrast to the random variation of SOC in plot A, there were high spatial dependences in the spatial distribution of SOC in the others. The spatial heterogeneity of TN in Picea meyeri forest plots were higher than those in subalpine meadow plots,and spatial autocorrelation were high in all the plots. The ranges of spatial autocorrelation variation (the distances within which parameters are spatially dependent) for SOC and TN were larger for subalpine meadow plots, but smaller for Picea meyeri forest plots.