OBJECTIVE The objective of this study was to assess the reliability of three-dimensional (3D) cephalometric approaches by comparing this method with authenticated traditional two-dimensional (2D) cephalometry in angular cephalometric measurements. METHODS CT images and lateral cephalometric radiographs of ten patients (five women, five men) were used in this study. Raw CT data of the patients were converted to 3D images with a 3D simulation program (Mimics 9.0, Leuven, Belgium). Lateral cephalometric radiographs were used manually for 2D measurements. The comparisons of the two methods were made using 14 cephalometric angular measurements. The Wilcoxon matched-pairs signed-ranks test (alpha = 0.05) was used to determine the difference between the two methods. To assess the intra- and interobserver reproducibility, two sets of recordings made by each observer, in each modality were used. Dahlberg's formula was used to determine the intraobserver reproducibility, and the Wilcoxon matched-pairs signed-rank test (alpha = 0.05) was used to assess the interobserver reproducibility. RESULTS The method errors of both observers ranged from 0.35 degrees to 0.65 degrees. In addition, there were no significant differences between the measurements of the two observers (P > 0.05). However, comparison of 2D and 3D parameters showed significant differences in U1-NA and U1-SN measurements (P < 0.05). CONCLUSIONS The 3D angular cephalometric analysis is a fairly reliable method, like the traditional 2D cephalometric analysis. Currently, the 3D system is likely to be more suitable for the diagnosis of cases with complex orthodontic anomalies. However, with the decrease in radiation exposure and costs in the future, 3D cephalometrics can be a suitable alternative method to 2D cephalometry.