Reliability is a fundamental attribute of the supply chain and is the basis for its proper operation. It is very meaningful to estimate the reliability function of a supply chain accurately. In this paper, we use the copula function to measure the dependence structure among suppliers and propose an optimal model averaging method based on Kullback-Leibler (KL) loss, which estimates the reliability function of the supply chain with a k-out-of-n: G system by weighting different copula functions. We prove the asymptotic optimality of the proposed estimator and the consistency of weights. Further, simulation studies and a real dataset demonstrate the effectiveness of the proposed method.