Vegetation and rainfall pattern are two important factors affecting the infiltration-runoff processes in slopes. Rainfall events in the eastern margin of the Tibetan Plateau, monitored from June 2019 to October 2020, were classified into four patterns using the self-organizing mapping method. To determine the relationship between infiltration potential, runoff generation, soil erosion, and rainfall parameters, the four rainfall patterns were simulated on the bare and vegetated slopes, and soil moisture content, surface runoff, subsurface flow, and sediment yield were measured. The results showed that rainfall characteristics and meteorological parameters significantly affected soil moisture supply. Rainfall intensity and soil moisture content mainly affected the initial infiltration rate, soil properties and root characteristics determined the stable infiltration rate. Vegetation could accelerate the rate and amplitude of soil moisture absorption and dehumidification. Soil erosion intensity and rainfall duration were found to follow a power function (R-2 > 0.84) in the bare slope and an exponential function (R-2 > 0.64) in the vegetated slope. Vegetation could significantly mitigate soil erosion by altering the hydrodynamic parameters of overland flow, and the mitigating effect was found to be more pronounced for light rainfall events. Moreover, the existence of an optimal initial moisture content for minimizing soil erodibility could be ascertained. These results provide insight into the dynamic responses of infiltration potential and erosion process in bare and vegetated slopes to different rainfall patterns and could guide the prevention of slope instability and soil erosion in response to complex climate change in the next two decades.