A skeletonization algorithm of foreign fibers in lint cotton was proposed. Distance transformation was implemented to the original image and the distance map was obtained. The local maxima points of the distance map were treated as the seed points and the skeleton was extended forward in all directions along the ridge points. Then the local maxima points were connected through the ridge lines. The skeleton points set scan and skeletonthinning process were carried on to eliminate the duplicated points, which had no use to the skeleton connectivity. Then the pruning process was implemented to remove the pseudo-skeletons and the accurate skeletons were obtained. The proposed algorithm guaranteed the connectivity of the skeleton and single-pixel in width. Furthermore, it could extract accurate skeletons of foreign fibers with slender shape and discontinuous parts. Test showed that the proposed algorithm had low computing complexity and could achieve the fast extraction of foreign fiber skeletons. 