Wind farms provide a source of clean and renewable energy. However, unlike many industries where machines are operated under more or less static conditions, wind turbines suffer from stochastic loading due to the hourly or seasonal variation of wind speed and direction. The stochastic loading of wind turbines makes their degradation or failure prediction rather complex. This in turn makes the decision-making process of when and what type of maintenance action to undertake very challenging. This paper uses the discrete event system specification (DEVS) to develop a simulation model for wind farm operations and maintenance. The DEVS methodology1 provides a formal modeling and simulation framework based on dynamical systems theory and allows for hierarchical and modular model construction. We report on implementation results based on historical data that provide useful insights into wind farm operations under two different maintenance strategies, scheduled maintenance and condition-based maintenance. The results show that condition-based maintenance enables more wind power generation by reducing wind turbine failure rates and thus increasing wind turbine available.