Target tracking has become an area of interest in recent years. Target tracking is able to widen the perspective and view of the static camera to track a target, providing basic artificial intelligence features for robots, and serve as the platform for surveillance purpose. A tracking system is limited to a monocular vision which limits much of the information that is available in an image compared to a binocular or stereo vision. Among the advantages of stereo vision is 3D modeling of scene and depth estimation. In this work, an initial development study and implementation of stereo vision tracking system was done. In developing the stereo vision tracking system, a tracking algorithm was developed together with a depth estimation method to take advantage of the images obtained from having images from two cameras. The depth estimation was found to be 80% accurate for objects at a distance of 15 cm to 65 cm from the cameras.