Adaptive website can automatically improve its hyperlink structure by learning from diverse web user access patterns. however, existing methods might change the website topology to excessive, which has already been familiar to must users. as navigation information contained in original topology changed, this might cause misleading navigation to users. therefore, a tradeooff solution between website adaptation and navigation information maintenance is necessary. in this paper, we firstly present the formal description for measuring the adaptation impact on navigation information, and then introduce a multi-objective optimization method of adaptive website (MOMAW) for navigation information maintenance based on genetic algorithm. we experimented by simulating users browsing website and compare MOMAW against adaptation without any optimization. our results show that MOMW can not only do gradual adaptation along with continuously emerging user access patterns, but also maintain more consistent navigation information.