Background: When we make treatment plan of class III malocclusion children, it is difficult to determine whether we treat it with orthognathic surgery or without orthognathic surgery. To determine that, we must consider many factors, such as cephalometric analysis, growth pattern, family history, and skeletal age. A Harvold cephalometric analysis is useful in determining the amount of discrepancy by comparing the maxillary unit length with mandibular, unit length. We tried this study to help the decision of treatment planning in class ill malocclusion children by comparison in class III malocclusion and normal occlusion children using a Harvold analysis. Materials and Methods: The materials for this study consisted of 20 class III malocclusion children. Cephalometric tracing and measurements were performed by one investigator. The control group consisted of 18 normal occlusion children and lateral cephalograms were obtained from 8.5 to 14.5 years old children biannually. The relationships between class III malocclusion group and normal occlusion group were evaluated statistically. Results: The lower anterior facial heights between two groups were not significantly different, although the lower anterior facial heights of class III malocclusion group was higher than those of normal occlusion group in all age groups. The Maxillary-mandibular unit length differences of class III malocclusion group were significantly higher than those of normal occlusion group(p<0.05). Conclusion: A Harvold analysis was useful to make treatment planning for class III malocclusion children.