The water quality from nonpoint source run off results from different land use types has been studied. The construction of a buffer strip is one method of nonpoint source pollutant control. The Soil and Water Assessment Tool (SWAT) model has been applied to estimate the pollutant removal through the buffer strip. When the non-business land has been changed into grass to form a buffer-strip, the change of land use effects the results of the model according to measures of the water quality. The data from a water level station within the watershed in the years 2006 and 2007 was used for calibration and validation of the model. Under the rainfall conditions in 2007, the removal rates of SS, BOD, TN, TP were 11.5%, 9.5%, 1.2%, and 4.5%, respectively. During the rainy days, the removal rates at the buffer strip were 92.3% of SS, 91.2% of BOD, 82.4% of TN, and 83.5% of TP. The pollutants from nonpoint sources were effectively removed by over 80% as they passed through the buffer strips. Rainfall resulted in soil erosion, which led to an increase in the SS concentration, therefore, the construction of buffer strips protected the streams from SS inflows. Since TN concentrations are affected by the inflows of ground water and the N concentration of the rainfall, the removal rate of TN was relatively lower than for the other pollutants.