Abstract: This article takes typical broad-leaved forests in Sichuan Province as the research object, and uses variance analysis and redundancy analysis to analyze the driving mechanisms of forest factors and terrain factors on the distribution of fuel load. Then, a stepwise regression method is used to construct a fuel load prediction model for eight typical broad-leaved forest types (evergreen oak, deciduous oak, hard broad-leaved, soft broad-leaved, Cinnamomum, Populus, Betula and Eucalyptus). The research results indicate that: there are significant differences in the composition of combustible load components among different forest types, and their distribution patterns are synergistically regulated by stand parameters such as canopy density and age structure, as well as terrain factors such as altitude and slope orientation. The total combustible load model constructed based on key influencing factors has high prediction accuracy, with a determination coefficient of greater than 0.707 for model fitting and testing, and a root mean square error of less than 10.4 t·hm⁻². This study provides a quantitative tool for precise estimation of forest combustibles and fire risk assessment, aiding in fire early warning and prevention.