With the aid of SWAT/GRASS interface, SWAT model was applied to a representative rural watershed and evaluated for its applicability through calibration and verification using observed data. Results of the model application showed that the simulated runoff was in a good agreement with the observed data and indicated reasonable applicability of the model. In predicting sediment yield, however, there were differences between the simulated and the observed values. It seems due to the inaccurate determination of MUSLE factors, better prediction of sediment yield could be achieved by selecting more appropriate MUSLE factors for different land use types. In predicting nutrient loading, the simulation results showed a similar trend to observed data.