The field scale experiment was performed to examine the performance of the constructed wetland for nonpoint source (NPS) pollution loading reduction. Four sets (each set of 0.88 ha) of wetland (0.8 ha) and pond (0.08 ha) systems were used. Water flowing into the Seokmoon estuarine reservoir from the Dangjin stream was pumped into wetland systems. Water depth was maintained at 0.3-0.5 m and hydraulic retention time was managed to about 2-5 days; emergent plants were allowed to grow in the wetland. The wetland effluent concentrations of $BOD_5$, TSS, and T-N were higher in winter than in the growing season excepting the T-P, and effluent $BOD_5$ concentration was higher than influents in winter. Mass retention of T-N and T-P was stable throughout the year, whereas mass retention of $BOD_5$ and TSS was decreased in winter. $BOD_5$, TSS, T-N, and T-P performance of the experi-mental system was compared with the existing database (North American Treatment Wetland Database), and was within the range of general system performance. From the first-order analysis, T-P was virtually not temperature dependent, and $BOD_5$ and TSS were more temperature dependent than T-N. Overall, the wetland system was found to be an adequate alternative for treating polluted stream water with stable removal efficiency and recommended as a NPS control measures.