Abstract：【Objective】Soil organic carbon (SOC) and soil organic carbon density (SOCD) are critical indicators of forest soil carbon sequestration capacity and play an essential role in regional and global carbon cycling. Mountain forest ecosystems are characterized by strong environmental heterogeneity and pronounced vertical differentiation of soil properties, this leads to complex spatial patterns and depth-dependent driving mechanisms of soil carbonpools. However, the spatial differentiation and underlying controls of SOC and SOCD in Picea schrenkiana forests of the central Tianshan Mountains remain poorly quantified. Therefore, the objective of this study was to reveal the spatial distribution patterns of SOC and SOCD across different soil layers and to clarify their dominant driving mechanisms along the soil profile.【Method】This study was conducted in Picea schrenkiana forests of the central Tianshan Mountains based on 267 sampling plots and 801 soil profile sampling points. Soil samples were collected from two depth intervals (0–20 cm and 20–60 cm) to determine SOC content, SOCD, soil bulk density (BD), soil nutrients, nitrogen pool characteristics, and soil moisture. Geostatistical analysis was applied to characterize the spatial distribution patterns and spatial autocorrelation of SOC and SOCD across the study area. Variance partitioning analysis (VPA) was used to quantify the relative contributions of soil properties, topographic factors, and vegetation attributes to the spatial variability of SOC and SOCD in different soil layers. Furthermore, structural equation modeling (SEM) was employed to disentangle the direct and indirect effects of key environmental variables and to explore depth-dependent changes in the driving mechanisms of SOC and SOCD. 【Result】For all sampled points, SOC and SOCD showed a significant decreasing trend with increasing soil depth. In the surface soil layer (0–20 cm), SOC and SOCD exhibited similar spatial distribution patterns, with relatively high values in the western and northern regions and relatively low values in the southeastern region. In contrast, in the deeper soil layer (20–60 cm), the relatively high-value areas of SOCD shifted toward the southeastern region, while the western and northern regions became relatively lower, indicating a clear spatial centroid shift compared with SOC distribution. The dominant factors controlling SOC and SOCD differed markedly between soil layers. In the surface layer, SOC was primarily driven by labile nutrients, especially alkali-hydrolyzable nitrogen and available potassium, reflecting the strong influence of biological and biochemical processes. SOCD in the surface layer was jointly controlled by SOC content and soil BD. In the deeper layer, SOC depended more strongly on nitrogen pool size and soil moisture conditions and was strongly dominated by soil BD. VPA results showed that soil factors explained 45%–87% of the spatial variation in SOC and SOCD, which was significantly higher than the explanatory power of topographic and vegetation factors, and the contribution of soil factors increased with soil depth. SEM analysis further confirmed a pronounced shift in driving mechanisms along the soil profile. Surface SOC and SOCD followed a “labile nutrient-driven SOC accumulation–SOCD enhancement” biochemical pathway, whereas deep soil carbon storage transitioned toward a mechanism driven by “nitrogen pool size plus physical protection (BD).” Deep-layer SOCD was highly sensitive to soil BD, and soil moisture indirectly reduced BD, resulting in a “physical structure trade-off” effect.【Conclusion】This study demonstrates clear vertical differentiation and spatial heterogeneity of SOC and SOCD in Picea schrenkiana forest soils of the central Tianshan Mountains. The results highlight a depth-dependent transition from nutrient-driven biochemical controls in surface soils to physically dominated mechanisms in deeper layers. These findings enhance the understanding of micro-scale vertical regulation of forest soil carbon pools in arid mountain ecosystems and provide a solid scientific basis for improving regional carbon cycle models and implementing targeted soil carbon sequestration and management strategies.