Understanding the effect of vegetation restoration on soil carbon sequestration is one of the important pathways to evaluate the performance of ecological restoration efforts, and is valuable in estimating the potential of humans for climate change adaptation. In the 1980s and late 1990s, owing to the Ecological Immigrants Project and Grain for Green Project, wide spread conversions of cropland to grassland and forest have happened in the karst region in Southwest China. Several studies have been carried out to determine the changes of vegetation carbon storage induced by land cover change in this region. However, because of the high inherent terrain variability of the karst landscape, relatively little is known about the accurate size of the current soil organic carbon (SOC) storage and the degree of human-induced changes. Therefore, the objectives of this study were to develop an appropriate method for estimating SOC storage in a typical karst landscape at a local scale and to estimate SOC change responding to vegetation restoration due to the Grain for Green Project. For this study, a 200 m×200 m grid was established over the study site (10.24 km²) and a total of 249 surface soil samples (0-15 cm) were collected in 2011, and 81 soil profiles were investigated by vegetation type (cropland, forage, plantation, regressed land, shrub land, secondary forest) in 2009 and 2011. The profiles of the soil samples were separated into depth segments:0-10, 10-20, 20-30, 30-50, 50-70 cm, and 70-100 cm. The maximum sampling depth was not greater than 100 cm and soil depth less than 100 cm was sampled to bedrock. A total of 424 samples from 81 profiles were collected. Vegetation type, canopy cover, slope, aspect, soil depth, and rock exposure (4 m×4 m) were recorded for each sample point. Additionally, soil depth and rock exposure were measured at 150 randomly selected points in the study area. All data were transformed for normality of distribution and homogeneity of variance prior to analysis. Pearson correlation test was performed to determine the relationships between SOC and various environmental parameters. Cokriging interpolation of the parameters of negative exponential profile depth function was utilized to estimate soil carbon storage in the study area. The results showed that, regional SOC storage can be estimated accurately by cokriging interpolation of logarithmic parameters fitting for carbon density and depth of each soil profiles, and the prediction performed well with a determination coefficient of 0.723. Conversion of cropland to forage, and regressed land significantly increased the SOC content from 19.3 g/kg in cropland to 23.5 g/kg in forage, and 34.6 g/kg in regressed land, whereas it slightly decreased when transforming the cropland to plantation (17.8 g/kg). Overall, vegetation restoration by the Grain for Green Project increased soil carbon storage for 23.43% in the study area. Additionally, transformation of cropland to forage is a superior approach for both ecosystem restoration and economic benefit in degraded karst ecosystems.