A new hybrid visual servoing based tracking of multiple targets using a swarm of mobile robots is proposed. This distributed algorithm has position based visual servoing (PBVS) and image based visual servoing (IBVS). In addition, the proposed method consists of two approaches: interaction locally among robots and target tracking. Furthermore, neural network extended Kalman filter (NEKF) is used for reducing noises which is existed during tracking targets. When the targets are slower than the robots, Lyapunov function can be used for showing that the robots asymptotically converge to each vertex of the desired configurations meanwhile tracking the targets. Towards the algorithm practical execution, it is necessary to identify the observation ability of each robot in an efficient and inexpensive way. Infrared proximity sensors and monocular camera are applied to fulfill these requirements. Our simulation results describe the proposed algorithm confirms that the considered distributed tracking multi-targets method applying robots swarm is effective and straightforward to implement.