Potential runoff and erosion is a serious problem for some types of sprinkler irrigation systems, particularlytraveling laterals and center pivots on medium to heavytextured soils operating on sloping land. Prediction of when runoffmight occur is part of the system design process. The USDAARS Water Erosion Prediction Project (WEPP) model was testedwith 3 years of field data under highrate sprinklers in southern Idaho. Runoff and soil loss were measured on the upper,middle, and lower portions of a hillslope. The main parameter affecting infiltration and runoff was the effective hydraulicconductivity. Model predictions for average runoff and soil loss were improved when hydraulic conductivity values wereadjusted to account for soil variability across the field. Runoff amounts were small, and prediction variability for individualfurrows was quite high, but no more than would be expected from previous studies of infiltration variability. Soil losspredictions were unreliable for the small runoff amounts occurring in this study. The most reasonable use of WEPP forsprinkler irrigation would be for estimating when potential runoff might occur under center pivots for different soils, slopes,and crop management practices, and to determine limits on application depths and rates to avoid serious runoff.