Abstract: The cost of multi-beam satellite transmitters is high, while the communication demand and communication priority of ground cells are always in dynamic change. This paper investigates a dynamic beam hopping strategy to achieve comprehensive area coverage and on-demand services under limited satellite resources and fewer transmitters. Unlike the traditional cluster beam hopping idea, in order to improve the spectral efficiency of the satellite system, global beam hopping is considered to be applied over the whole frequency bandwidth in this paper. Therefore, a new quality of service metric is considered to measure the beam hopping results in the context of co-channel interference. In the case of limited satellite resources, the communication capacity provided by each beam cannot meet the service requests of ground cells. An adaptive beam power allocation algorithm is proposed based on the particle swarm optimization algorithm for the beam hopping pattern. The algorithm minimizes the beam capacity requirement gap by weighted priority, and improves the quality of service level. Finally, the performance superiority of the algorithm is verified by simulation.