[Background]Soil erodibility K value is a required parameter of soil erosion models, it is an index to indicate the sensitivity degree of soil erosion, and an accurate K is the prerequisite for constructing soil erosion model. Studying soil erodibility K-factor of granite collapse region contributes to macro-scope estimation and quantitative analysis on spatial variation characteristics of different soil layers. [Methods] Collecting eluvial horizon, illuvial horizon and parent material horizon of different soil layers in granite collapse region developing in Tongcheng, Hubei Province, then estimating by using five estimation methods of soil erodibility K value (nomo equation, modified-nomo equation, EPIC model, Shirazi model, Torri model), artificial simulated rainfall experiments were conducted to verify the effectiveness of the Soil erodibility K values of different soil layers in granite collapse region and the sensitivity of five estimation methods. According to the situation of the annual rainfall, topography and geomorphology, the rainfall intensity was designed (70+4) mm/ h, rainfall duration was 40 min, the slope was 20°. [Results] 1) Parent material horizon (PMH) was mainly composed of sand particles; clay content was the lowest with the average of 8.61%. The particles in illuvial horizon (IH) were mainly silt and sand, and the soil organic matter content difference was significant. The mass fraction of organic matter in eluvial horizon (EH) was 1.24%, far higher than that in IH and PMH, therefore, EH possessed the strongest anti-erosion ability, the second for IH, and the worst for PMH. Thus, in the governance of collapse mound, the PMH and IH should be protected. Deng Liangji, et al [24] has revealed that in development process of collage mound, each layer of soil erodibility K value will change, therefore the change of K value was investigated as below. 2) The soil erodibility of different soil layers in granite collapse region were significantly different, the average K value of the parent material layer was the highest, it was 1.20 times of the eluvial horizon and 1.03 times of the illuvial horizon; the stable sediment rate and the loss of particle sizes of different soil layers in granite collapse region were also significantly different. Erodibility K value of different soil layers by nomo equation was the closest to the stable sediment rate of different soil layers at 40 min precipitation. The sensitivity in estimating the soil erodibility K value of different layers by nomo equation was the highest, it was 1.5 times of modified- nomo equation and 6 times of EPIC model. [Conculsions] Therefore, nomo equation can accurately evaluate soil erodibility value of different soil layers in the granite collapse region. By estimating soil erodibility K value and the spatial variation characteristics of different soil layers in granite collapse region, this work is of certain guiding significance for the particular study of the formation mechanism and its governance of granite gully.