Abstract: In view of the characteristics of wireless sensor networks (WSNs) and their high timeliness in target tracking, a tracking strategy based on dynamic clusters is proposed. When the target appears, analytic hierarchy process (AHP) is used to select a cluster head, and the tracking weight of a node is defined and used to judge whether the node should join in the cluster for tracking, based on the node management mechanism. During the tracking, the cluster activates suitable sensors to participate in the tracking task, and the range of the activation area is adaptively adjusted according to real-time tracking quality. Besides, an adaptive target recover mechanism is proposed to avoid target losing. After the target leaves, the cluster is deconstructed to release sensor resources. The NS2 simulation results indicate that the dynamic cluster strategy for target tracking can get high tracking rate at low cost of communication overhead.