Abstract: Currently, due to the increasing data amount in the Internet of Things, it is challenging to meet the the requirements for quality of business data services and achieve high bandwidth utilization rate. To address this problem, a multi-priority queuing theory bandwidth allocation method is proposed. First, the business data transmission process from perception terminal to edge IoT was improved. The improved transmission process enables prioritization of data transmission according to the different requirements of various data, and facilitates the implementation of different service mechanisms for different priority levels. Then, the Markov process was analyzed during business data transmission. Based on the improved data transmission process, a multi-priority bandwidth allocation model with the packet loss rate and the delay time taken as constrains was developed to optimize bandwidth utilization . The multi-priority queuing theory bandwidth allocation method is compared with traditional methods, resulting in an improved QoS indicator and a 9.73% higher bandwidth utilization rate compared to traditional bandwidth allocation methods. Additionally, it achieved a 31.17% higher utilization rate than the elastic coefficient method. Finally, the dynamic performance of the multi-priority queuing theory bandwidth allocation method was also explored. The results indicate that the proper improvement of the bandwidth can improve the QoS indicator; however, it should be noted that increasing the bandwidth may lead to a reduction in its utilization rate. Therefore, reasonable bandwidth allocation can avoid waste of resources effectively.