Cluster-based topology is an energy-efficient topology control method in cognitive radio networks. The cluster head node depletes energy faster than cluster member node and the rotation of cluster head node is needed to balance the energy consumption for the whole network. Aimed at the problem of unbalanced residual energy of each node caused by cluster heads and cluster members in wireless cognitive radio network, a novel algorithm named cluster head rotating threshold (CHRT) is presented. In the proposed algorithm, rotation energy threshold is estimated using cluster head real-time energy load. The simulation results show that comparing with LEACH and EDAC, CHRT can realize the more network lifetime than the others.