This paper introduces a new model for node behavior namely Correlated Node Behavior Model which is an extension of Node Behavior Model. The model adopts semi Markov process in continuous time which clusters the node that has correlation. The key parameter of the process is determined by five probabilistic parameters based on the Markovian model. Computed from the transition probabilities of the semi-Markov process, the node correlation impact on network survivability and resilience can be measure quantitatively. From the result, the quantitative analysis of correlated node behavior on the survivability is obtained through mathematical description, and the effectiveness and rationality of the proposed model are verified through numerical analysis. The analytical results show that the effect from correlated failure nodes on network survivability is much severer than other misbehaviors.