Abstract: Aiming at the problem of low performance and high energy consumption in the cloud computing system, this paper proposes a resource scheduling algorithm based on dynamic matching mechanism. Tasks were classified according to the value and emergency, and the global cloud task queue was divided into four sub queues by four quadrant laws. The availability evaluation model was established by using colored Petri Net with memory identifier, and the nodes were dynamically divided into four-level resource pools according to the different availability zone. The tasks in sub-queues were matched and scheduled to four-level resource pools, at the same time, the resource pool was managed by different power management technologies. The simulation results show that compared with the traditional resource scheduling strategy, the proposed strategy can effectively guarantee the user service performance and significantly reduce the total energy consumption of the system.