Field experiment was performed from August 1996 to December 1999 to examine the performance of constructed wetland system for wastewater treatment in rural area. The constructed wetland system was installed in Konkuk University and the effluent from septic tank of school building was used as an influent to the treatment basin. The treatment basin was composed of sand bed with planted reed. From August 1996 to June 1998 the hydraulic loading rate was fixed with about 15.63cm/day and theoretical detention time was 1.38 days, and from July 1998 to December 1999 the hydraulic loading rate was about 6.25cm/day and theoretical detention time was 3.5days. It worked continuously even during winter time, and the sewage flowed without freezing even when average daily air temperature was below -1$0^{circ}C$. Average removal rate of BOD , COD, and SS was about 70%, T-P removal rate was about 50.8% , and T-N removal rate was 23.9%. The reason for poor T-N removal might be due to high influent concentration and short retention times. At the later years BOD and COD removal rates were increased , and SS and T-P removal rates did not change significantly , but T-N removal rates were decreased. The effluent of the wetland system often effluent water quality standards for sewage treatment plant, therefore, further treatment would be required if the effluent need to be discharged to the public water. Wetland system involves relatively large land area and could be suitable for rural area. Therefore, utilization of reclaimed sewage for agricultural purpose or subsequent land treatment is recommended as a ultimate disposal of sewage for agricultural purpose or subsequent land treatment is recommended as a ultimate disposal of sewage in rural area.