Abstract: This paper established a preliminary run mechanism of adaptive cartographic visualization system, and particularly analyzed each module of system. Based on the above, the paper researched into the monitoring mechanism of user action, and used apriori association rules improved to the mining of user action database to adapt to ACViS. The experiment proved that the mining results have guiding meanings for ACViS designers and provide basis t for the renewal and modification of user model. The researches lay theoretical and technical foundations for follow-up research and development of adaptive cartographic visualization system.