Non-point sources pollution has become a serious environmental problem in the aquatic systems throughout the world. The Xitiaoxi catchment is located in the southwest of Taihu Basin, contributing large amounts of runoff and associated nutrients to Taihu Lake. Thus, identifying critical non-point sources pollution in this catchment is urgent and essential to control water pollution, improve the water quality, and reduce the pollutants drained into water bodies. The present study integrated a monthly water balance model with the export coefficient model for total nitrogen and total phosphorus loads estimation in the Xitiaoxi catchment in southeastern China. The simulated monthly runoff are in good agreement with the observed streamflow at both Hengtangcun and Fanjiancun stations with Nash–Sutcliffe coefficients higher than 0.80. The predictions showed reasonable ranges from 1687 to 2046 t/y (2002–2005) for total nitrogen loads, and from 106 to 157 t/y for total phosphorus loads (1999–2007), respectively, which are consistent with the observed values at Hengtangcun. Overall, the monthly export coefficient model coupling monthly water balance simulation to export coefficient model presented both the seasonal dynamics and magnitude for streamflow and nutrients loads, which generally match well with the observations. These findings demonstrate that the proposed model can provide encouraging results and can be used as an efficient tool to identify the pollution sources for planning and management of large-scale agricultural catchment.