Energy-harvesting technology enables sensor nodes to collect energy from the external environment, which greatly alleviates the energy consumption problem of wireless sensor network. In order to make full use of the collected energy and improve the network performance, a dynamic energy harvesting wireless sensor network routing algorithm is proposed. The algorithm can dynamically adjust the number of cluster heads in each round according to the average residual energy and collected energy of the network to ensure efficient use of collected energy and avoid node energy overflow or waste. In the cluster head selection, the residual energy, the collected energy and the distance between nodes are taken into consideration at the same time, so that the cluster head nodes with better energy status can be evenly distributed in the network. The simulation results show that the proposed algorithm can effectively balance cluster heads distribution, and is superior to other existing algorithms based on energy-harvesting model in terms of average energy efficiency and throughput.