Abstract: Soil loss in karst area is divided into surface soil loss and soil leakage. The binary and three-dimensional hydrogeologic structure in karst area provides space condition for its soil leakage. Different from surface soil loss, soil leakage is the transport and deposition of soil from the surface to the underground space. The strong karstification provides multifarious paths for soil leakage. Due to the complexity and variability of loss paths, the diversity and interactivity of influencing factors and the multi-interface nature of subsurface hydrological processes, a quantitative analysis of soil leakage in karst area is always one of the important issues in soil erosion research and one of the difficult questions in soil erosion forecasting and monitoring. Focused on influencing factors of soil leakage, this study is aimed at analyzing the effects of environmental factors such as soil properties, vegetation, rainfall characteristics, terrain feature and human activities on soil leakage. For quantitative tracing of sediment source in such levels as runoff plots, cave catchment area and watershed in spatial scale, four main quantitative methods about evaluating soil leakage in karst area, such as simulated runoff plot method, cave drip tracer method, traditional model method and composite fingerprinting, have been analyzed and compared. These four quantitative research methods have their own advantages and limitations. The simulated runoff plot method can quickly and intuitively monitor the soil leakage at a small spatial and temporal scale. However, its result is highly sensitive to external environmental factors such as rainfall and physical and chemical properties of soil. The cave drip method can only trace the soil loss that occurs by cave dropping water, which is quite different from the actual loss in the cave catchment. However, as a new leakage research method, it provides a new idea for leakage monitoring. The determination of soil leakage at watershed scale mainly includes model method and fingerprinting identification method. The traditional model method can directly monitor the leakage at the watershed scale, but there are some limitations in practice. First of all, the traditional method requires clear underground runoff outlet in the basin and no exchange and superposition between underground runoff and surface runoff. Secondly, it is requested that the surface and underground sediment production and drainage only occur in the basin without the disturbance by other basins. Thirdly, for the traditional model method, the underground or surface runoff sediment discharge should be monitored at fixed points, and the accuracy of sediment amount is greatly affected by the location of monitoring point and monitoring time. To some extent, fingerprinting identification method can be used to solve some problems of traditional model method, but it also has some limitations such as the selection of sediment sources, the screening of fingerprinting factors and the correction of retention of them. In this paper, the future study focuses on soil leakage in karst area are also pointed out by analyzing the problems of research on soil leakage. The collation of quantitative research on soil leakage in karst area provides a reference for exploring the driving mechanism of soil leakage and for further studying the coupling relation between soil leakage and environmental factors.