Abstract: This study analyzed spatial variation features and compared assessment methods of soil organic carbon (SOC) storage in the 015 cm layer in high bulk-ratio slope of Karst Regions. Three SOC storage assessment methods (soil patch sum method, classical geo-statistical method and discontinuous soil interpolation method based on bulk-rock ratio and soil layer depth) were analyzed in the study. Using grid sampling method, about 107 soil patch samples were set up in a 100 m × 100 m plot. The SOC content and soil bulk density averages were respectively 75.5 g·kg^-1 and 0.8 g·cm^-3 with relatively moderate coefficients of variation of 30.6%, and 47.3%, respectively. The structural characteristics of SOC semivariogram followed an exponential model with nugget, sill and range values of 260.8, 521.7 and 52.5 m, respectively. Semivariograms of SOC content and soil bulk density showed strong anisotropy when the lag distance ranges were 015.2 m and 34.754.2 m, respectively, implying that micro-topography and topography significantly influenced structural variations in SOC. Based on soil patch sum assessment method, SOC storage of the plot and SOC density were respectively 983.8 kg and 0.1 kg·m^-2. Also for discontinuous soil interpolation method, the estimated SOC storage of the plot and SOC density were respectively 2 712.8 kg and 0.3 kg·m^-2. The estimated carbon stock of the plot and SOC density by the classical geo-statistical method were 86 264.0 kg and 8.6 kg·m^-2, the carbon stock was 87.7 times of the actual value, with estimation error rate of 8 668.4%. The results suggested that the classical geo-statistical method was not suitable for estimating carbon stock and density in high bulk-ratio slope of Karst Regions. Compared with the classical geo-statistical method, revision of the bulk-stock ratio and soil depth significantly reduced estimation error of carbon stock and density by the discontinuous soil interpolation method. Therefore discontinuous soil interpolation method was more suitable for the estimation of soil carbon stock and density in such areas than the classical geo-statistical method. It was concluded that although the geo-statistical was an effective approach for analyzing spatial heterogeneity of SOC, it failed to precisely assess spatial distributions of soil patches, habitat types, bulk-rock ratios and soil thicknesses. Thus there was the need to revise the estimation methods in order to get the actual value of SOC. This study provided the scientific basis for future studies on the assessment of carbon storage in Karst Regions.