There are usually several potential failure modes in mechanical components. The conventional model for the system analysis is built under the condition that all failure modes are independent of each other. However, in engineering practice, failure modes are mostly dependent due to the fact that the elements involved in each failure mode are closely interrelated. System reliability analysis and evaluation simply conducted under independent assumption often result in excessive errors or even wrong conclusion. A novel method to evaluate system reliability of mechanical components with multiple failure modes based on moment method are proposed here. Firstly, the moment-based reliability and reliability sensitivity analysis method is proposed with independent assumption. Secondly, the proposed method is deduced by taking the correlation between failure modes into account and the correlation model is established with the copula function, which is proved to be a useful tool to model nonlinear correlation with marginal distributions. The robust design is performed as a biobjection optimization process based on the reliability sensitivity. The numerical examples show that the applied procedure is able to efficiently consider various failure modes of mechanical components in probabilistic assessment and reliability-based optimization.