Optimal reactive power dispatch (ORPD) is a particular case of the optimal power flow. ORPD was usually considered as the minimization of an objective function representing the total active power losses in the electrical networks. The constraints involved are the generator voltages, tap regulating transformers ratios and the amount of reactive shunt compensators. The goal of this study is to find the best vector of control variables, so that the power loss decreasing can be realized. In this paper, a new metaheuristic technique inspired from the bubble-net hunting technique of humpback whales, namely whale optimization algorithm (WOA), has been applied to solve the ORPD problem. The WOA method has been examined and confirmed on the IEEE 14-bus, IEEE 30-bus, in addition to a practical and large scale Algerian electric 114-bus test system. The obtained outcomes have been compared with two own developed methods, namely, particle swarm optimization (PSO) and particle swarm optimization with time varying acceleration coefficients (PSO-TVAC). Afterwards, an analysis of variance (One-way ANOVA test) has been implemented in order to verify the performance of our proposed algorithm in solving the ORPD problem. In summary, the comparison study shows the potential of this recent optimization method, and proves its robustness and effectiveness in solving the ORPD problem.