Mobile Target Tracking is viewed as one of the primary applications in Wireless Sensor Networks (WSNs). Due to WSNs' special characteristics, there happens to be a trade-off between tracking accuracy and energy consumption. We propose an efficient multi-target tracking method in which the tracking task is switched between static clusters as the backbone of the network and dynamic clusters, which form in boundary regions. To reduce energy dissipation in case of extra dynamic cluster formations, we apply a merging mechanism where intersecting dynamic clusters appear. The numerical results achieved using the Castalia simulator verify the efficiency of the proposed scheme comparing to HCTT algorithm, with 10% decrease in the average power consumption and an average of 36.4% decrease in the number of redundant data packet transmissions at the cost of 1.5% increase in the control packets' overhead. Satisfying results are also obtained for tracking quality of our method compared to a prediction-based method, DPT.