Abstract. Current Semantic Web technologies provide a logic-based framework for the development of advanced, adaptive applications based on ontologies. But the experience in using them has shown that, in some cases, it would be convenient to extend its logic support to handle vagueness and imprecision in some way. In this paper, the role of vagueness in the description of Web user interface characteristics is addressed, from the viewpoint of the design of adaptive behaviors that are connected to such descriptions. Concretely, vague descriptions combined with quantified fuzzy rules and flexible connectors are described, and their usefulness is illustrated through preference modeling, filtering and adaptive linking scenarios.