Abstract: To solve the problem that the relationship between various types of damage and different influencing factors of bridge under earthquake is complicated and difficult to find, the damage situation of the bridge was analyzed combined with bridge pier type, structure type, bearing type, plane line shape, foundation type, intersection angle, fault distance, and peak ground acceleration. The conditional attribute of bridge damage were reduced by using the dominance relation-based rough set approach. The minimum sets of association rules of 374 bridges in Wenchuan earthquake were extracted. Another 20 bridges were selected for testing. Results were compared with indiscernibility relational rough set theory and classification and regression tree algorithm. Results show that for the evaluation system of post-earthquake bridge damage state, compared to the traditional generation rule based on indiscernibility relation rough set theory and classification and regression tree algorithm, the dominance relation-based rough set approach is more accurate, and can be used in the future research of bridge damage discrimination studies and rule extraction.