Mobile Edge Computing (MEC) is one of the most promising paradigms for overcoming Edge Devices (EDs) constraints. These EDs suffer from resource limitations in terms of power and computation.MEC will be more prevalent with the rising resource-intensive and time-sensitive EDs applications. MEC is considered a superior alternative to cloud computing. Despite computational offloading to the cloud offeringsignificant benefits related to computing and storage, EDs are geographically distant from the cloud, leading to significant transmission delays. However, offloading to the nearest server and ignoring the huge capabilities of the cloud is not always a good option. In contrast, local computing is rarely preferable. On the other hand, sometimes offloading to the nearest server is impossible, because of the current state of the server. These possibilities, as well as MEC system unpredictability, make the offloading decision difficult and critical. Therefore, the idea of the proposed model is based on Reinforcement Learning (RL). Moreover, the model is designed to make an optimal decision amongthe three offloading options; nearest edge server, best edge server, and cloud. The edge server can decide to offload tasks to the optimal available edge server or cloud directly, which depends on several parameters for reducing execution time and energy consumption. In addition, the edge server connects to all componentswithin its region, which improve the managing of resource allocation. This proposed model is expected to be optimal in edge servers connection and intelligent offloading decisions. © 2022 IEEE.