[Objective] The spatial database of geological disaster field survey was conducted to analyze the formation mechanism and failure mode of collapse in the area where granite weathered residual soil was developed, to explore the contribution rate of different environmental disaster factors to collapse development, and to evaluate the spatial susceptibility and distribution law of collapse disaster.[Methods] Based on the analysis of the development characteristics of granite weathered residual soil collapse at Baiyun District, Guangzhou City, eight disaster-causing factors closely related to the development of collapse disaster (including elevation, slope, slope direction, topographic relief, distance to stream, rainfall, stratum lithology, and land use) were selected to construct the evaluation index system of collapse disaster susceptibility, and 404 historical collapses in the study area were superimposed. Using the contribution rate model, the collapse disaster sensitivity and spatial distribution characteristics of each index factor were calculated and counted. The modified sample difference of the factor contribution rate of each secondary index was used to determine the factor weight.[Results] ① The contributions of slope, elevation, and topographic relief to collapse susceptibility in the study area were relatively high, while slope direction, stratum lithology and rainfall had low contributions to the assessments of collapse susceptibility. ② The extremely high-prone areas were mainly concentrated in the western foothills, and the collapse susceptibility decreased gradually from the outer edge of the mountain ground line to the center. ③ The collapse rates of the extremely high and high prone areas accounted for more than 85% of the total collapse rate. The model susceptibility evaluation success rate and prediction rate reached 91.3% and 92.6% respectively.[Conclusion] Topography and geomorphic factors of slope had significant impacts on collapse development. The collapse disaster susceptibility evaluation model based on contribution rate could objectively quantify the index factor weight, and the prediction results of the model had high accuracy. The susceptibility zoning conformed to the actual spatial distribution characteristics of collapse development.