Incomplete ordered information processing is a common problem in the real life. Various extended dominance relation rough set models are already proposed to solve the incomplete ordinal decision problem. But there exits some ambivalence over the real semantics because the characteristics of the order relation are not considered. Therefore, the confidence dominance relation and its rough set approximation model are presented. Relationships between the confidential dominance relation based rough set model and existing models are discussed. Compared with the current expansion relations, confidence dominance relation fits the characteristics of the order relation to avoid semantic contradiction. Theorem proving and example analysis show that the confidential dominance relation based rough approximation model has better approximate accuracy and classification accuracy.