A road sign detection, recognition, tracking algorithm based on multi-cues hybrid is presented. In detection stage, the color cue is used to segment the interesting regions, then the corner and geometry cues are used to detect the signs. In recognition stage, a coarse classification is preformed by using the corresponding relationship of color and shape, then the support vector machine with binary tree architecture is built to recognize each category of road sign. In order to reduce recognition errors, Lucas-Kanade feature tracker is introduced for road sign tracking. The experimental result shows that the proposed approach has better robustness.