A system of systems (SoS) is a class of complex systems that views multiple and small independent systems as part of a larger and more complex system. Because designing an SoS requires evaluation from a long-term viewpoint, it creates a large uncertainty. While researchers have proposed some robust design methods to handle the uncertainty of systems, they cannot be adopted to SoS design because of the SoS’s strong nonlinearity. This paper first discusses how difficult it is to handle uncertainty of SoS considering the SoS characteristics, and then proposes an approach of robust optimality with lattice points that is simple but plausible to sidestep the issue of nonlinearity. A case study of designing a distributed generation (DG) system, which is a typical SoS, in a Japanese dormitory town is demonstrated in order to verify the proposed approach. The results reveal that this approach helps a designer decide the appropriate amount of subsidy in terms of the robustness of DG installation.