Product reliability and manufacturing cost are two essential factors in increasing competition within industries. Most studies on predict product or system reliability are based on failure rate models, and assume components of reliability system are independent of one another. The compatibility among components is thus ignored, making predictions of the system reliability imprecise. This study focusing primarily on a non-repairable compound series-parallel system, determines the optimal parameter settings for each component using the dual response surface method (DRSM), simultaneously considering the system reliability and manufacturing cost. The Box-Behnken Design (BBD) from response surface methodology (RSM) is initially used to produce the design, and the experimental data, including system reliability and manufacturing cost, are gathered using Monte Carlo Simulation. The optimal parameter setting associated with system components is determined to obtain a highly reliable and robust system, using DRSM, which is applied to maximize the system’s reliability subject to particular manufacturing cost. Accordingly, the significance of interaction effects is evaluated to elucidate the compatibility among components. The proposed approach can not only accurately predict the system reliability, but also let customer requirements be incorporated into the reliability system, and reduce substantially the time taken to develop of a new product.