Empirical soil loss models are relatively simple and have been applied in many parts of the world. The application of the Soil Loss Estimation Model for Southern Africa (SLEMSA), the Universal Soil Loss Equation (USLE) and the Revised Universal Soil Loss Equation (RUSLE) in southern Africa is investigated. Strictly speaking, these empirical models are modelling techniques or frameworks, rather than mechanistic descriptions of the system and should, therefore, make no claim of universality. The current set of sub-models for calculating SLEMSA factors were developed specifically for the Zimbabwean highveld, but the sub-models are being used to give best estimates of sheet erosion throughout Zimbabwe and elsewhere in southern Africa. As an index of the spatial distribution of soil loss, SLEMSA can be regarded as a useful model to differentiate areas of high and low erosion potential. USLE is the most widely known and used soil loss equation in the world. In southern Africa efforts were made to determine local values for USLE factors, but the model was never widely implemented. RUSLE can be described as a software version of a greatly improved USLE. A few studies, executed under different conditions in southern Africa, showed its potential for use outside its country of origin. The model proved to be a promising tool for conservation planning, as it provides a dynamic approach to predicting soil loss. The need is recognised,however, forimproved verification of the results. A theoretical evaluation and sensitivity analysis performed on the above three models, clearly showed the advantage of themoreflexible and dynamic structure of RUSLE against the strict empirical structures of USLE and SLEMSA. Although empirical models could be expected to be used mainly as screening tools in integrated studies, land resource assessments would demand increased accuracy in quantification of erosion rates in a spatial and temporal context. If data requirements can be met, process-based models (e.g. the Water Erosion Prediction Project, WEPP) can be expected to find increased application in producing quantitative estimates of soil loss and sediment yield in southern Africa.