In view of the wireless sensor network used for emergency situations, this article presents a dynamic clustering based on event-driven dynamic clustering data fusion algorithm. On the basis of considering the severity of the incident and the residual energy, the algorithm chosen stimulus intensity and number of nodes remaining energy cluster-head play. Cluster-head node cluster by broadcasting a message of dynamic clusters, and dynamically adjust the delay time of each round of data fusion, data fusion to improve efficiency. Simulation results show that compared with the TEEN protocol, this algorithm has a lower average energy consumption of nodes and a better integration of network data latency.