Multi-agent systems are used to solve problems that are difficult to be solved by an individual agent or a monolithic system. Many multi-agent approaches are adopted to develop vast and complex systems in different application areas. Interactions among agents are usually ruled by interaction protocols that function to ensure the correct system execution. The basic function of the interaction protocols is to allow the agents to communicate effectively without needs to have planned arrangements or saving space for each speech act communication messages. The agent communication language ACL has proven success in software based multi-agent systems, but it suffers from constraints that make it impractical for transferring low level of data. The interaction protocol is more efficient with a small size of information needs to be transmitted in a short communication time. Hence, the efficiency of the interaction protocol is low when transferring a big size of information for a long time of communication. FIPA English and Dutch auction interaction protocols are not addressing the problems of cancellation, refuse, and failure as they have limited communication acts between the customers and the auctioneer, which are restricted for "propose" and "inform" messages only, unlike the rest of the protocols that addressing these problems. This research will investigate the performance of FIPA English and Dutch auction interaction protocols for a multi-agent system in order to propose a sort of recommendations that would contribute to enhance the performance and the efficiency of the FIPA interaction protocol, through examining the performance of the interaction model before and after applying a communication acts of (Cancel, Refuse, and Failure) within the communication acts between the auctioneer and the customers. Different levels of credibility (Low, Moderate, and High) are applied on the FIPA English auction interaction protocol (FIPA-EAIP). And it has been concluded that the lower the interactions number, the better the performance of the entire system, as the system will spend less time and power in communication, which would lead to enhance the performance and the efficiency of the entire system and network. KEYWORDS: Multi-agent system, Agent Communication Language, Interaction Protocol, FIPA, English Auction, Dutch Auction.