Abstract. In this paper, we introduce the concept of rule-based visualization for a computational steering collaboratory and show how these rules can be used to steer the behaviors of visualization subsystems. Feature-based visualization allows users to extract regions of interests, and then visualize, track and quantify the evolution of these features. Rules define high level user policies and are used to automatically select and tune the appropriate visualization technique based on application requirements and available computing/network resources. Such an automated management of the visualization subsystem can significantly improve the effectiveness of computational steering collaboratories in wide area Grid environments.