Abstract: In the industrial Internet environment, edge computing has the advantage of low communication latency, but it cannot cope well with the demand of multi-device multi-tasking on latency. To address the above problems, we proposed to combine edge computing with cloud computing, built a computation offloading model based on cloud-edge-end collaboration, and introduced an M/M/S queuing model in order to alleviate the computation pressure of edge computing in multi-device multi-tasking scenarios. According to the task queuing, a reasonable task offloading was performed, and a computation offloading strategy based on MS-SSUA was proposed. The experimental results show that the offloading utility of this strategy is stable and more flexible compared with other offloading strategies and reduces the cost of task unloading delay.