This paper is concerned with prognostic information for maintenance and production optimization in complex systems. In each stage of such a system, we allow redundant components used as backup to ensure system's availability. Remaining Useful Life (RUL, the so-called prognostic information) of components is used to evaluate each component's redundancy. We address RUL-based maintenance and production optimization to guarantee the availability and productivity of the system such that client demands can be satisfied in a given planning horizon. We propose a mixed-integer linear programming model to minimize the total cost. Experimental results on test instances show the efficiency of the proposed approach to attain optimal solutions.