The optimal placement of distributed generation in power distribution systems problem consists of defining the most appropriate sites to install those generators and their optimal sizing, aiming to optimize the system performance. From the mathematical point of view, it is a complex nonlinear optimization problem, containing continuous and discrete variables. The present paper deals with the optimal placement and sizing of distributed generators for real power losses minimization in distribution systems. For this purpose, a new metaheuristic approach called War Optimization is proposed. The best sites to place these generators are defined by the War Optimization method, and the sizing is given by an optimal power flow tool. A set of simulations is run using radial distribution test systems containing 69 and 476 busbars. A detailed comparison between War Optimization and other metaheuristics (including Particle Swarm Optimization) shows that the proposed method is efficient, as it presents better solutions more often.