Summary: We consider a decentralized control design for a class of fuzzy complex systems. This system contains uncertainty, consists of multiple subsystems, and faces information structure constraints. The uncertainty is represented by an uncertain bound, which is described by a fuzzy membership function. The control scheme is decentralized. In addition, it is robust and optimal: while it guarantees certain performance regardless of the uncertainty, it also minimizes a performance index. With regard to the minimization phase of the design, the solution is shown to be existent and unique.