This work relates to context-awareness of things that belong to IoT networks. Preferences understood as a priority in selection are considered, and dynamic preference models for such systems are built. Preference models are based on formal logic, and they are built on-the-fly by software agents observing the behavior of users/inhabitants, and gathering knowledge about preferences expressed in terms of logical specifications. A 3-level structure of agents has been introduced to support IoT inference. These agents cooperate with each other basing on the graph representation of the system knowledge. An example of such a system is presented.