Energy efficiency is one of the most challenging issues in wireless sensor networks, particularly in the target tracking. The main purpose of these networks is to preserve the distributed important sites from the targets who intend to destroy the sites in an environment. This paper proposes a distributed energy-efficient mechanism for sleep scheduling the sensors followed by a dynamic cellular clustering algorithm for tracking the targets. Probabilistic positions of the targets are predicted based on an improved particle filter procedure consecutively. A cell, including a subset of sensors, is constructed over time considering the predicted positions of targets. This cellular algorithm not only decreases the number of awake sensors, but also increases the tracking quality. Moreover, a concept named communication base has been proposed for alleviating the communication volume of the sensors. Utilizing communication bases, sensors are able to exchange information in maximum three hops. The experimental results demonstrate the capability of the provided algorithms and mechanism in optimizing the energy of sensors and increasing the tracking quality.