The impacts of topsoil stripping on soil erosion and sediment yield and their characteristics were analyzed in karst slope by the simulated rainfall experiments. The “dualistic structure” of karst slope was simulated by a steel tank filled with soil and rock. An orthogonal test was conducted to study the characteristics of underground runoff yield and sediment yield in different underlying surface types (slope farmland, bare slope), different rainfall intensities (0.5, 0.8, 1.3 mm/min) and different degrees of underground hole (crack) (1%, 2%, 3%, 4%, 5%). The results showed that the changes of soil surface physical properties caused by topsoil stripping were the imperative affecting factors on the underground runoff and sediment yield variations. Underground runoff and sediment yield in slope farmland were mainly impacted by rainfall intensities, while the distribution ratios of runoff and sediment in the surface and underground were mainly influenced by different degrees of underground hole (crack). Topsoil stripping resulted in a significant increase in soil bulk density and fractal dimension, which made the soil water permeability, pore size and other soil physical properties change. When the rainfall intensity was 0.5mm/min, the underground runoff yield, runoff coefficient and sediment proportion in the bare slope were greater than those in the slope land, but when the rain intensity increased to 0.8 and 1.3 mm/min, the results were just reverse, the corresponding values in slope land were greater than the bare slope. The runoff and sediment yields in slope land with different rainfall intensities dwere 0.5 > 0.8 > 1.3 mm/min. However, the runoff yields in bare slope were 0.8 > 1.3 > 0.5 mm/min, and the sediment yields were 0.5 > 0.8 > 1.3 mm/min. The increased degree of underground pore fissure had a significantly (p<0.05) effect on increasing contribution of runoff and sediment yield of slope. Under the degree of underground hole (crack) of 1%~3%, there was a significantly difference (p<0.01) between the underground erosion proportions in two slopes, but the difference diminished?gradually with the degree of underground hole (crack) increasing to 4% and 5%. It was worth noting that, although the amount of underground sediment yield of the bare slope was less than that of the slope land, the proportion of underground sediment yield of the bar slope was greater. It should be paid more attention to control the underground leakage where farmland was employed to development and construction in karst area. The results could provide great assistance and reference for further research and controlling water and soil erosion in the Karst region.