Traffic congestion and the search for parking spaces have become significant challenges in modern urban development. These issues lead to increased fuel consumption and contribute to stressful lifestyles. To address this, efficient parking allocation has emerged as a challenge in urban environments, requiring innovative solutions to optimize the use of available parking spaces. Various approaches have been employed to tackle this problem, including mathematical programming and heuristic procedures, which typically follow a centralized model. Another promising solution involves the use of multi-agent systems (MASs), which adopt a decentralized approach. However, there is limited research on the application of MASs in parking allocation. This paper presents a multi-agent system where intelligent agents represent vehicles and parking spaces to efficiently allocate parking spaces to cars. The paper evaluates four coordination mechanisms based on the Contract Net Protocol for parking allocation. The results show that the concurrent use of the Contract Net Protocol with decommitment capacities is the most effective approach, as it reduces the time required for parking space allocation while optimizing the utilization of available parking spaces.