In the current networking field, research works are going on with the aim of providing user%s desired QoS everywhere. This paper discusses the distributed multi agent scheme (DMAS) scheme used in the IMS-4G-cloud networks to support session (service) continuity. It explains the role of different agents which make use of knowledge sources in solving the problem and performing tasks. A Q-learning awareness algorithm is used to estimate the QoS by calculating cost function with the help of mathematical formulae comprising of QoS parameters like jitter, delay, packet loss, network resource availability, mobility and service fare parameter. Agents in the DMAS interact with each other and work cooperatively to provide both QoS and mobility information. Based on the type of information provided, two types of handoff came in to existence. A detailed explanation of these handoffs with phases is discussed in this paper. Finally, a comparison is made between general IMS handoff procedure and DMAS-IMS handoff procedure with the help of signal flow diagram.