The purpose of this study was to show the method of three-dimensional morphometry developed recently and to compare the accuracy of three-dimensional morphometry with those of PA cephalometry, The three-dimensional morphometry analysis program and device were developed. Steel balls (1.2mm in diameter) were attached in twenty five landmarks of artificial human skull. This artificial human skull was used as experimental materials. From three-dimensional morphometry and PA cephalometry of artificial human skull. eleven linear measurements were acquired and made into asymmetry index. Right-left differences of measurements were used as asymmetry index. These measurements and asymmetry index were compared respectively with those of actual. The results were as follows: 1. Mean difference between three-dimensional morphometry and actual artificial human skull in linear measurements was $1.99{pm}0.37mm$, and mean difference between PA cephalometry and actual was $21.12{pm}0.45mm$. Both of all were reduced more than those of actual. 2. Mean difference between three-dimensional morphometry and actual artificial human skull in asymmetry index was $0.07{pm}0.42$, and mean difference between PA cephalometry and actual was $3.63{pm}0.60$. Three-dimensional morphometry was reduced while PA cephalometry was magnified more than that of actual. 3. Each eleven asymmetry index of three-dimensional morphometry was the same negative sign as those of actuals while only N-Z, ANS-J, Tr-Go, Tr-ANS asymmetry index were the same in PA cephalometry. These results suggest that the method of three-dimensional morphometry is more accurate than those of PA cephalometry in asymmetry analysis.