Due to the fact that the effect of node blocking on traditional task dynamic scheduling results in poor emergency task dynamic scheduling, this paper proposes a data center emergency task dynamic scheduling method based on queue migration. It builds an emergency task dynamic scheduling model, analyzes the cloud platform task dynamic scheduling process according to the model, and completes the optimization operation from the background, so that the emergency task request is sent to the appropriate queue of the dynamic scheduling center for dynamic scheduling, thus obtaining the best mapping from the emergency task to the processing unit. It analyzes the time characteristics of data processing emergency tasks, solves the processing blocking problem of each node, and determines the effective time and waiting time of data processing tasks. Combined with controlling the queue migration process of dynamic scheduling nodes, the dynamic scheduling process of emergency tasks in data center is designed. Experimental results show that this method can complete all tasks with the completion rate of more than 95% and the cut-off missing rate of always less than 18%, and it has a good dynamic scheduling effect.