Abstract: In ordor to study the distribution of forest soil organic carbon(SOC), a total of 840 soil sampling sites were laid based on a grid of 4 km×6 km, using portable GPS. Normality test was applied for log-transformed data of forest SOC data. Moran’s I, geostatistics and GIS techniques were used to study the spatial variation and spatial patterns of SOC. Extremely significant spatial auto-correlation of forest SOC was found based on the global Moran’s I value. Local Moran’s I index revealed that high SOC values were located around the area of Qiandao Lake, the west part of Lishui city, while low SOC values were located in the middle part of Zhejiang province. The experimental variogram of SOC was calculated and fitted by exponential model. Further ordinary Kriging was applied to interpolate the SOC in forests of Zhejiang province. The spatial patterns of SOC were strongly related to human being activity such as planting economic forest, changing soil management and so on.