The performance of the EUROSEM model is evaluated for a catchment in the Netherlands. The model is calibrated using observed hydrographs and sedigraphs by interactively changing the input parameters net capillary drive, initial soil moisture content, saturated hydraulic conductivity, cohesion, Manning's n and the detachability of soil particles. Storms with characteristics similar to the calibration storms were simulated well, whereas the model did not perform too well for storms being significantly different. Simulation of soil loss was generally too high though the poor quality of observed soil loss made it difficult to assess the true performance of the model. Through a plus and minus 10% sensitivity analysis, it was revealed that the model was most sensitive to changes in the initial volumetric moisture content of the soil and Manning's n and at the same time, the sensitivity analysis proved to be both condition- and site-specific in nature. Because of the difficulties of parameterising the model, it is recommended that in future applications of the model, minimum and maximum output values should be included as the output. Improved within storm modelling is needed, particularly to model conditions where crusting is widespread in the study area.