Abstract: Landmarks of Cephalogram is the important basis to diagnose and treat patients with facial skeletal and dental abnormalities. This paper presents an approach for automatically locating landmarks on cephalograms by applying image processing and artificial intelligence techniques. This approach first uses filtering and edge detecting techniques to extract edges of soft and hard issues in cephalograms. Then it recognizes the landmarks on the basis of the anatomical knowledge of the human facial structure. This approach not only realizes the automatically landmarking but also makes the diagnosis quantitative. The accuracy of the approach in locating the landmarks is compared with values obtained from human experts.