In this chapter, the independence between failure modes is relaxed and copula models are then used to model dependent failure modes in one-shot devices under constant-stress accelerated life-tests. The chapter describes a composite likelihood approach in detail for the likelihood estimation of model parameters with the use of two prominent Archimedean copulas, namely, Gumbel–Hougaard and Frank copulas. Copula models have recently become quite popular in reliability studies for capturing dependence between lifetime variables corresponding to different parts or components. Along the lines of Ling et al., an extensive Monte Carlo simulation study is conducted to examine the performance of the proposed estimation method for the Gumbel–Hougaard and Frank copulas with weak and moderate dependence, with Weibull and gamma lifetime distributions. The chapter considers serial sacrifice data with R codes.