As the penetration level of wind farms in power system increases, studying their impact on the different aspects of power system such as reliability will be necessary. Various factors such as wind speed patterns, wind turbine technology and wind turbine arrangement in wind farm, influence power output and reliability of wind farms. When wind blows through a wind farm with multiple rows of wind turbine generators (WTGs), the downstream wind speed after a WTG is lower than the upstream wind speed. This phenomenon is called the wake effect. Wind farms are usually arranged regularly in some rows and columns because of different reasons such as land constraints, ease of access and maintenance and simpler cabling system and collection grid. Wake effect of upstream wind turbines reduces power output of downstream wind generators, so this phenomena should be considered in calculating reliability indices of wind farm. Wake effect in most of references related to reliability of wind farms, has not been considered. Some references considered this phenomena using simulation methods such as Monte Carlo. These methods usually take much time to response and they need long-term wind speed data that can not be easily prepared in a particular wind farm. In this research, the Markovian Frequency and Duration Method as an analytical method, is developed to consider the wake effect of regularly arranged wind farm similar to a multi-state power plant. This model also takes into account Fixed Outage Rate (FOR) of wind turbines and calculates probability, number of occurrence and frequency of each state of the wind farm model. We used sparse matrix techniques to solve computational problems due to dimensionality of large wind farms. Some of the advantages of this model are: good response speed compared to simulation methods, no need to long-term wind speed data and compatibility with reliability assessment methods used for ordinary power plants. Then, by using the proposed model, the system adequacy indices such as LOLE, LOLF, LOLD and EENS are evaluated and the impact of wake effect is analysed. A Modified version of the test system of RBTS is used and the accuracy and efficiency of the proposed method is verified. Keywords: Wind farm, Power system reliability, Markovian Approach, Wake effect