The maritime environment is diverse, open but partly ruled, and covers a large spectrum of ships from small sailing boats to tankers which generally exhibit type-related behaviours. Several real-time positioning systems have been designed for keeping track of ships movements. However data provided by these reporting systems are underutilized to help operators in charge of maritime safety and security. This paper introduces an information system relying on user-defined navigation rules and an inference engine to analyse maritime traffic and report abnormalities and suspicious behaviours to operators. The proposed approach improves existing ones by combining a spatio-temporal database and a rule-based inference engine.