In this paper, an analytical approach for the reliability modeling of large wind farms is presented. A systematic method based on frequency and duration approach is utilized to model a wind farm like a multistate conventional unit, where the probability, frequency of occurrence, and departure rate of each state can be obtained using the regional wind regime of wind farm and wind turbine characteristics. The proposed method is capable of finding both annual frequency and average time of load curtailment analytically in the presence of wind power. A wind farm in the northern region of Iran with the wind speed registration of one year is studied in this paper. To accommodate time-varying patterns of wind speed, reliability analysis considering seasonal patterns of wind speed is also carried out. The results show that seasonal patterns significantly affect the reliability indexes. A reliability analysis is also performed using a load profile similar to that of Iran power network. It will be shown that the coincidence of high-load-demand and high-wind-speed periods makes the North Iran wind farm projects highly attractive from a reliability point of view.