An investigation on inter-annual variations in wind resource and the accuracy of WAsP prediction was carried out in a real wind farm located at Hangwon, Jeju island, Korea. The wind data were obtained from an anemometer, and the wind vane installed on the meteorological mast and the operational data of the wind turbines tested were given by a monitoring system. The data were obtained for each six-month period in 2005 and 2006, respectively. As a result, it is clear that the wind characteristics vary year by year. Also, it was found that the cup anemometer on the meteorological mast in Hangwon wind farm was affected by the wake behind wind turbines. Considering the wake effect, the accuracy of WAsP prediction for the power production was assessed. The result shows that the relative error for the corrected power production that considered the wake loss was less than ${pm}10%$.