Mineral-bound organic carbon in soil, especially iron-bound organic carbon (Fe-OC), plays a pivotal role in organic carbon storage. Understanding the controlling factors and distribution patterns of Fe-OC is crucial for improving Earth system model predictability in the study of land C-climate feedbacks. This study, centered on the Tibetan Plateau, involves soil sample collection from various locations. Chemical analysis and machine learning methods are employed to unveil the relative significance of climate and soil factors in Fe-OC formation. Results indicate that soil organic carbon (SOC) content predominantly influences Fe-OC formation, with climate factors exerting a comparatively minor impact. Among soil factors, pH, SOC content, and cation exchange capacity are identified as decisive in Fe-OC formation. By training random forest models and analyzing variable importance, the spatial distribution of Fe-OC on the Tibetan Plateau is successfully predicted, with concentrations primarily ranging between 0.95-5 g/kg. Ultimately, this research reveals Fe-OC content and distribution across various grassland types and soil depths on the Tibetan Plateau. High-altitude cold grasslands on the plateau significantly impact global carbon balance, and the estimated Fe-OC reservoir provides valuable references for future global carbon cycle studies.