This paper presents a modeling methodology for designing networked agents systems. It focuses on the agent hybrid state representation, its dynamic behavior, and its interaction with neighbor agents. The agent’s internal state as well as the communication interface between agents is modeled using timed hybrid Petri nets.