Abstract: Soil physicochemical properties serve as crucial driving factors for carbon storage. As a vital ecological barrier zone in China’s southwestern border region, the Sichuan-Yunnan ecological barrier zone has received limited research attention regarding the specific influences of soil physicochemical properties on carbon storage dynamics. This study integrated the InVEST model with the optimal parameters geographical detector (OPGD) and multiscale geographically weighted regression (MGWR) model, drawing on multi-source datasets—including land use/land cover data at 30-m resolution, soil properties at 1-km resolution, and climate variables—to systematically examine the spatiotemporal evolution patterns of carbon storage and the spatially heterogeneous driving factors in the Sichuan-Yunnan ecological barrier zone from 2000 to 2020. The results are summarized as follows: 1) Over the study period, total carbon storage exhibited an overall declining trend, which was primarily attributable to rapid urban expansion and intensive agricultural development in lowland river valley regions. These anthropogenic activities drove widespread conversion of woodland and grassland into construction land and cropland, resulting in substantial losses in aboveground and belowground biomass carbon pools, as well as accelerated soil organic carbon release due to tillage disturbance and erosion. In contrast, localized increases in carbon storage were observed in certain areas, largely benefiting from woodland expansion and large-scale ecological restoration projects such as the Grain for Green Program (returning farmland to forest), which enhanced vegetation cover, litter input, and soil organic matter accumulation. 2) Normalized difference vegetation index (NDVI), soil organic matter content (OC), and digital elevation model (DEM) emerged as the dominant driving factors, each with the explanatory power exceeding 50%. All multi-factor interactions were of the enhancement type, with the strongest synergistic effect observed between NDVI and OC (NDVI ∩ OC). This finding underscored the dominant role of the coupled ecology–terrain–soil nutrient mechanism in shaping the spatial distribution of carbon storage across the zone. 3) The MGWR model successfully captured significant spatial heterogeneity in the effects of driving factors, confirming that soil physicochemical properties played a pivotal role in influencing both the distribution patterns and transfer processes of carbon storage, particularly in areas with strong topographic gradients. These findings provide practical guidance for future land use optimization, ecological governance, and carbon sink management in the Sichuan-Yunnan ecological barrier zone. By highlighting the central importance of soil properties, vegetation–soil interactions, and spatial heterogeneity, this study offers a scientific foundation and an innovative multi-model analytical framework for sustainable carbon storage management in similar ecological barrier zones. Ultimately, the results contribute to climate change mitigation efforts, support enhanced regional ecological conservation, and promote ecosystem resilience enhancement under ongoing human and climatic pressures.