Blackouts in the operation of power systems generate economic losses that increase in proportion to the time it takes to recover the electricity supply. For this reason, it is necessary to develop restoration plans that guarantee secure operating conditions, facilitate decision-making by the operators of the power system and consequently minimize the blackout time. In this sense, this paper proposes a methodology according to optimization approaches based on constructive heuristic algorithms to determine the skeleton transmission network in a reconfiguration problem focused on restoration process. The proposed methodology determines the skeleton network avoiding overloads of the transmission elements into the scheduling of load pickup and couples islanding systems establishing the generation/load balance. The proposed methodology was applied to IEEE test systems and a realistic system corresponding to a mining system without or with black-start units, so islanding systems can be defined. The methodology finds feasible solutions very close to the global optimum despite using a constructive heuristic algorithm; this characteristic is very attractive in restoration planning due to its flexibility and easy implementation in bulk power systems.