Users' preferences have traditionally been exploited in query personalization to better serve their information needs. Most of the time, user preferences depend on context, that is, they may have different values depending on the situation of the user. In this paper, we propose a fuzzy-rule-based model for the representation of contextual preferences in a database querying framework. We discuss the augmentation of a query with preferences deduced from information regarding the current context of the user. To this end, we make use of an appropriate inference pattern, called generalized modus ponens, able to deal with data and knowledge when they are described in a fuzzy way.