Modern distributed information systems, among many applied technologies, rely on multi-agent systems, which is due to the characteristic properties of the latter - autonomy, mobility, intelligence and cooperation. These properties, providing power, flexibility and adaptability in solving many distributed, dynamic, fuzzy problems, are gaining in importance. As in many cases, the features of these properties require a special analysis for them while ensuring the operability and reliability of the functioning of distributed information systems, especially in critical applications. The considered model of verification of the dynamic coordination property - one of the basic properties of agents’ cooperation - is based on extended Petri nets and automata behavioral experiments of check and recognition. The model for verifying the coordination of agent groups is distinguished by the behavioral check of cause-and-effect coordination-synchronization of conditions, events, actions, functions, both agents and the resource environment of their placement with the individuality and independence of agent goals. The verification model of agent coordination is designed to determine the conditions for the check and recognition of group agent interactions in the methods of testing of distributed information systems with agent properties and allows to reduce the computational complexity of behavioral analysis, check, recognition and diagnosis in general.