This is a review of research related to soil-erosion prediction technology. The trend in erosion prediction technology in the USA, Australia, and Europe is toward the development of process-based simulation models. The emphasis in erosion research on strictly empirically based models, such as the Universal Soil Loss Equation, is declining. With the process-based technologies come a new array of research needs. In the USA, The USDA Water Erosion Prediction Project (WEPP) has produced a new generation of soil-erosion prediction technology based on fundamentals of hydrologic and erosion science. The development of the new computer-based WEPP erosion model for estimation of rill and interrill erosion required an extensive review and analysis of current knowledge of the science of soil erosion by water. Research needs were identified. The relative importances of identified research needs were evaluated using a sensitivity analysis of the WEPP Model, which identified the more important of the input variables required to execute the model. The review of research reported here, along with a discussion of associated current needs for research, addresses four general areas: (i) fundamental erosion relationships, (ii) soil and plant parameters related to erosion, (iii) data bases, user interfaces, and conservation system design, and (iv) erosion model development and analysis.