Currently, many wind farms exist throughout the world and, in some cases, supply a significant portion of energy to networks. However, numerous uncertainties remain with respect to the amount of energy generated by wind turbines and other sophisticated operational aspects, such as voltage and reactive power management, which requires further development and consideration. To fix the problem of poor reactive power compensation in wind farms, optimal capacitor placement has been proposed in existing wind farms as a simple and relatively inexpensive method. However, the use of induction generators, transformers, and additional capacitors represent potential problems for the harmonics of a system and therefore must be taken into account at wind farms. The optimal location and size of capacitors at buses of an 80-MW wind farm were determined according to modelled wind speed, system equivalent circuits, and harmonics in order to minimize energy losses, optimize reactive power and reduce the management costs. The discrete version of the lightning search algorithm (DLSA) is a powerful and flexible nature-inspired optimization technique that was developed and implemented herein for optimal capacitor placement in wind farms. The obtained results are compared with the results of the genetic algorithm (GA) and the discrete harmony search algorithm (DHSA).