Recent technological advances enable realization of wireless sensor networks suitable for many applications including environmental monitoring, biological and chemical contamination detection, earthquake management. Depending on the application, the properties of sensor network databases include a distributed environment, approximate and long-running queries, large volumes of data, uncertainty and fuzziness in data and in queries, activeness, etc., which bring out new database processing and management problems. In particular, reactive sensor network database applications must be able to detect the occurrences of specific events or changes in the network state, and to respond by automatically executing the appropriate application logic. For such applications, an ECA rule-based fuzzy active database approach fits well. In this study, we propose an active database approach employing a fuzzy Petri net model, which processes uncertain sensor data and handles flexible (i.e., fuzzy and approximate) and continuous queries. The proposed model provides a natural data processing and management approach for sensor and actuator network applications that require reactive behavior.