<p>Total organic carbon in lake surface sediments (STOC) constitutes a pivotal biogeochemical component within the critical interface of the global carbon cycle, where recently deposited organic matter undergoes burial or mineralization, thereby playing a key role in carbon turnover and storage in inland water ecosystems. However, previous studies, constrained by local-scale observations and linear analytical methods, have not systematically characterized the global spatial patterns of STOC values and the mechanisms driving their variability. Here, by integrating STOC measurements from 89 lakes across the Tibetan Plateau with STOC data of 1070 lakes compiled based on a literature search, we present the first machine-learning-based assessment of global lake STOC patterns. Our results indicate that lake STOC concentrations peaked between 45°N and 70°N in the Northern Hemisphere, particularly in northern North America, whereas very high latitudes (&gt;70°N) and mid- to low-latitude regions exhibited comparatively lower values. Structural equation modeling revealed that among basin characteristics, soil organic carbon and ice/snow cover were the strongest direct controls on lake STOC distribution, while temperature indirectly influenced STOC by modulating both variables. Although human activities affect STOC globally, their overall influence was modest. We conclude that temperature primarily governs the global spatial distribution of lake STOC concentrations through its influence on catchment processes.</p>