Faced with environmental legislation imposed by authorities, manufacturers must review their short-and mid-long term strategies in order to integrate the environmental dimension into the decision making process. In that context, in this paper, we address the problem of an unreliable manufacturing system producing one product family type to meet a demand. We consider that the manufacturing system's operations generate harmful emissions into the environment. Hence, in addition to inventory, production and backlog costs, an environmental penalty is imposed when the emission level reaches a specific limit (cap approach). Because technology investments in production clean processes is often heavy and to improve its environmental strategy, we consider that demand can be partially satisfied by an unreliable subcontractor. The subcontractor which is characterized by a high development and innovation on sustainable technology requires a high selling cost and a random availability. This work examines this decision making problem in order to propose a new joint production and subcontracting control policy which takes into account the emission level. The objective is to optimize the total cost, which includes the inventory, backlog, production, emission and subcontracting costs, over an infinite horizon. An experimental approach combining simulation, experimental design and response surface methodology is used to solve the problem. Through numerical examples and further sensitivity analysis, the structure of the control policy is confirmed and analyzed. Thanks to the practical usefulness of the resolution approach, we provide a decision support system to help managers in deciding whether or not to subcontract, depending on green subcontractor characteristics (availability and cost).