Abstract: Background Karst region is one of the most vulnerable ecosystems in China, and the soil resources in this region are limited because of the slow pedogenesis processes of severe soil erosion influenced by the unique karst process. In recent decades, soil erosion resulted in large rocky desertification, reductions of agricultural productivity, and decreases of land fertility. Yellow soil is the major soil type in this region. It is essential to study sediment yield characteristics at slope scales for effective soil slope loss control, water and soil resources utilization and management. Methods Based on the data of field plots in Bijie of Guizhou Province in Southwest Cina and Ansai of Shaanxi Province in Northuest China, the characteristics and relationships of rainfall, runoff and sediment yield under natural rainfall events were investigated and analyzed by statistical methods, and the differences between two stations were compared and discussed. Results 1) Runoff and sediment yield in the yellow soil slope of Bijie mainly resulted from heavy rains and storms, and they were smaller than those of loess slopes in Ansai, and more evenly distributed. 2) Runoff was more relevant to precipitation than to rainfall intensity, which was contrary to the relationships on loess slope. As precipitation increased from A rainfall type (0 < P ≤ mm) to B type (P > 25 mm), the relationship between runoff and rainfall characteristics on yellow soil slopes changed from random to a linear function. While under the same change of rainfall type, the relationship between runoff and rainfall characteristics remained linear on loess slopes. The relationship of rainfall intensity and sediment yield on yellow soil slopes was in an exponential function, while it was in power on loess slopes. 3) When lacking process data, the equation M = 0.54e ^{0.11 p} can be used to predict soil loss on yellow soil slope for individual rainfall events. 4) Rainfall erosivity was linear to both runoff and sediment yield on both yellow soil and loess slopes, and it was a proper indicator to predict runoff and sediment yields on loess slopes, while it was not that suitable for yellow soil slopes. Runoff and sediment yield caused by unit rainfall erosivity on yellow soil slopes were smaller than those on loess slopes. 5) Sediment yield of both yellow soil and loess slopes were in linear function to the runoff. When runoff was similar, sediment yield on yellow soil slopes was smaller to that on loess slopes. Conclusions It is valuable to study soil erosion factors for quantitative estimation of soil erosion, especially in Karst region. Precipitation is a dominant factor that influences runoff and sediment yield on yellow soil slopes. The more precipitation, the more runoff and sediment yield. The amount of runoff and sediment yield on yellow soil slopes are smaller than those on loess slopes, however, the soil erosion situation in the Karst region is even severer due to the scarcity of soil resources here.