The optimal reactive power dispatch (ORPD) is a complex optimization problem involving discrete and continuous variables. The main objective of the ORPD consists on adjusting generator voltage set points, transformer taps and reactive compensations for reducing power losses, improving voltage profile or a given voltage stability index. Due to the no linear, non convex and multi-modal nature of the ORPD most of the techniques applied to its solution are based on metaheuristic optimization. This paper presents a literature review of the different metaheuristic approaches applied to the ORPD. The review focused on journal papers published in IEEEexplore and ScienceDirect in the last 15 years. It was found that classical metaheuristic techniques such as genetic algorithms and their variations have been frequently used to solve the ORPD. However, new techniques such as gray wolf optimizer, shuffled frog leaping algorithm and gravitational search algorithms have been recently adapted to approach the ORPD.