Microgrids have become an attractive strategy to guaranty a reliable, efficient and sustainable electricity delivery to the distribution network, as well as to offer effectively ancillary services to the power utility. Therefore, special features have to be introduced to the planning methodology, while the optimization problem has also to be re-evaluated to contemplate the large number of variables, multiple objectives, and different constraints and uncertainties. Hence, a probabilistic multi-objective microgrid planning methodology based on the optimal size and location of distributed generation is proposed. In this methodology, the mismatch power in islanded mode and the annual energy losses in grid-connected mode are minimized, while the residual power for reserve provision is maximized. For that purpose, probabilistic models, distributed storage units, and constructive constraints are contemplated in the methodology. The methodology is tested in an adapted PG&E 69-bus distribution system and the NSGAII optimization method is used to solve the optimization problem.