In this paper, we have developed an integrated maintenance policy for a manufacturing system subjected to a random failure and calling up on several subcontractor machines in order to satisfy a constant demand. In order to reduce the machine failure a preventive maintenance plan is made. The machine is unable to satisfy a constant demand, that's why it called upon another machine, comprising the so-called subcontractor machine, which produces at a certain rate the same type of product. In order to assure an economical objective, we have a choice between several subcontractors having some different data and to optimize the preventive maintenance plan. Two strategies are developed and optimized in this work. The first strategy, single subcontractor strategy, consists in choosing only one subcontractor among several ones. For this strategy we have proved analytically, that the choice of the subcontractor machines is conditioned by the unit lost cost due an unsatisfied demand of one product; The second strategy, switching strategy, consists in relaying on one subcontractor machine and then switching to another. An analytical study is developed in order to optimize firstly the preventive maintenance plan and secondly to determine the optimal switching date. A numerical example is developed for very strategy in order to apply the analytical results.