In many wireless sensor networks (WSNs), target tracking is an essential application. This paper studies the real-time target tracking algorithm and the implementation for a multi-target real-time tracking system. The system consists of a wireless sensor network which includes several distributed ultrasonic sensor nodes and a monitoring base station, and two robots as moving targets. To avoid the conflicts in the network, a sensor node task scheduling scheme, and an adaptive clustering and inter-cluster negotiation network protocol (ACICN) are proposed for the system. To cope with distributed and asynchronous measurements, data synchronization and Extended Kalman Filter (EKF) location algorithm are studied for the system. The experiments show that the system can effectively track multi targets simultaneously.