In order to determine the optimal set of maintenance strategies for a network of bridges with concrete decks, a maintenance strategy optimization method using interactive genetic algorithm is presented in this paper. Evolution is leaded into optimal comprehensive performance of multiple objects by way of inputting inaccurate preference information and fitness Sharing technique is introduced to improve the diversity of populations. Ultimately, the set of solutions satisfying different preferences can be obtained. Thus, selecting difficulty of optimal maintenance strategy can be greatly reduced. Data obtained from bridge inspection report of 2005 of Liaoning province are used to demonstrate the feasibility and capability of the proposed approach.