An optimal structural design method is proposed to determine the structures of energy supply systems in consideration of their multi-period operation. The optimization problem is formulated as a large scale mixed integer linear programming (MILP) problem with binary variables for selection and on/off status of operation of equipment and continuous variables for capacities and load allocation of equipment. The dependence of the performance characteristics of equipment on their capacities and part loads, as well as that of capital costs of equipment on their capacities, are incorporated into the optimization model. A decomposition method of solving MILP problems with the block angular structure is applied to derive a suboptimal solution close to the optimal one in a short computation time. The meaning and validity of the decomposition method are clarified through an analytical study on an energy supply system with a simple structure. The validity and effectiveness of the decomposition method are also clarified in terms of solution optimality and computation time through a. numerical study on another energy supply system with a complex structure.