With the increase in cases of using Bluetooth devices used in the electric vehicle charging systems, security issu es are also raised. Although various technical efforts have beed made to enhance security of bluetooth technology, v arious attack methods exist. In this paper, we propose an intelligent Bluetooth intrusion detection system based on a well-known machine learning method, Hidden Markov Model, for the purpose of detecting intelligently representativ e Bluetooth attack methods. The proposed approach combines packet types of H4, which is bluetooth transport layer protocol, and the transport directions of the packet firstly to represent the behavior of current traffic, and uses the t emporal deployment of these combined types as the final input features for detecting attacks in real time as well as accurate detection. We construct the experimental environment for the data acquisition and analysis the performance of the proposed system against obtained data set.