An indoor mobile robot must have the ability to follow and track a moving target-object. This paper proposes an effective target-following scheme for mobile robots. The positions of the mobile agents are acquired from an RFID-based localization system. To estimate the state of the mobile agents and to predict the trajectory, a state model and a Kalman-filter are used. The objective of this study is to generate a stable tracking path for the mobile robot. We proposed a Virtual Tracking-goal Algorithm (VTA) based on the predicted position of the target-object for this objective.