In this paper, a distributed estimation algorithm is constructed by using local information for each mobile follower-subsystem to keep the same pace with one leader-subsystem in the multi-agent system. The algorithm is used to "observe" the unmeasurable state, i.e., the velocity, of the leader. With the help of common Lyapunov function method, it is shown that each follower can track the active leader while the interconnection topology is time-varying. Finally, a numerical simulation is given for validation of the algorithm.