In order to reveal the spatial distribution characteristics and influencing factors of soil erodibility (K-factor) on the Loess Plateau, this research estimated the values of K-factor on the Loess Plateau based on EPIC model, geometric mean grain size model and Torri model. The impacts of soil physical and chemical properties, elevation, slope, and other factors on the spatial distribution of K-factor were compared using geographic detectors. The findings indicated that: (1) The mean values of K-factor on the Loess plateau estimated by EPIC model, geometric mean grain size model, and Torri model were 0.036, 0.034 and 0.041, respectively. The dominant soils on the Loess Plateau exhibited medium erodibility and medium-high erodibility, with significant differences in the estimated K-factor values among the models (F=4.460, p<0.01). (2) The K-factor of Loess Plateau showed significant spatial heterogeneity, with higher K-factor in the east and southwest and lower K-factor in the northwest. There were statistical differences in the proportion of medium erodible and medium-high erodible areas among different provinces. (3) Soil physicochemical properties (sand content,silt content, clay content, gravel content, bulk density, pH, cation exchange, basic saturation, exchangeable salt group, carbonate, sulfate, exchangeable sodium salt, electrical conductivity), elevation, slope, and slope direction showed highly significant effect on K-factor (p<0.01). The influence of soil physical and chemical properties on the spatial distribution of K-factor was stronger than those of elevation, slope, and slope direction, and the interaction between factors had more influence on K-factor than individual factors. The study could provide some theoretical basis for the accurate prevention and soil water erosion controlling and soil improvement on the Loess Plateau.