Abstract: The tracking system algorithm applied for markerless augmented reality applications based on FAST keypoints was presented. A small number of key frames of the real environment were calibrated offline and used to detect the target object in video frame. During the online stage, the target object was automatically detected by matching the FAST keypoints extracted on the input image and the selected key frames. The RANSAC algorithm was employed to further optimize point correspondences and discard outliers. The camera pose was accurately estimated using an image registration algorithm. Experimental results demonstrate that our algorithm was real-time, accurate, robust, and promote the AR application on the intelligent terminals.