Abstarct: With the expansion of the Internet, virtual space and the entrance of Internet of Things equipment into daily life, the number of active Internet users has increased significantly. Despite the expansion of cloud computing as a reliable platform for users' computations, due to the very high number of nodes connected to the Internet in different regions and the ever-increasing volume of information, bandwidth and network traffic, cloud computing no longer meets the quality and demands of users. The quality and speed of provided services have very important impact in satisfying users. However, the construction of cloud farms in different regions can lead to higher energy consumption. Today, managing approaches for connecting users to local servers, which are called edge computing, has increased significantly. By creating local administrations, spreading and distributing servers as edge servers in different regions, while managing bandwidth and placing computing services closer to the user locations, the efficiency of resources increases and network traffic decreases. Various requirements such as quality of service, decreasing of network traffic, reduction of energy consumption, load balancing in a fair manner and baxsed on available resources are among the things that are considered in resource allocation and management. In this research in order to optimize the process of resource allocation in edge of network, a new approach is proposed with the aim of predicting the behavior of users in the long run and allocating the edge server to workload on a fixed basis. The proposed approach considers the edge servers and optimize the allocation of users to servers while managing energy resources and reducing energy consumption in servers in the short time. In the long time, it also balances the computing pressure on the servers. Also, this approach employs the hidden Markov model to predict the behavior pattern of users and genetic algorithm in order to obtain an acceptable and near optimal solution for problem of edge server allocation to users. This approach can measure various parameters such as network load and energy consumption with different or the same weights simultaneously. To evaluate this approach, it was investigated in allocating edge servers to users using real data sets and the results show that the proposed approach can reduce energy consumption in the long term.