Abstract: At present, only the remaining energy of nodes is considered when electing the leader node in existing chain routing algorithms. Although EEPB adopts the remaining energy of nodes and the distance between nodes and the base station(BS) to determine which node is qualified to be the leader node, simply weighting these two factors does not completely eliminate the inconsistency of the two decisions. Accordingly, a chain routing algorithm based on evidence theory called CRET is proposed. In the leader node election stage, CRET utilizes two evaluation indexes which are the remaining energy of nodes and the distance between nodes and the BS to confirm which node is chosen as the leader of the chain by Dempster-Shafer (D-S) evidence theory. In the CRET algorithm, two membership functions are respectively established to describe two evaluation indexes, and the basic probability assignment functions of the two indexes are acquired. Then the combination discipline of D-S evidence theory is used to determine the final results of two evaluation indexes. In the chain construction stage, in order to avoid the generation of long chain, the node which has joined the chain is taken into account, and each one is connected to its nearest node. Simulation results show that CRET has superior performance over EEPB on balancing energy consumption among the nodes and extending the network lifetime.