PURPOSE: This study compared bone ages measured by the Greulich-Pyle (GP) and Tanner-Whitehouse 3 (TW3 ) methods and investigated the differences in predicted adult heights measured by Bayley-Pinneau (BP) and TW3 methods. METHODS: Bone ages were assessed from left-wrist radiographs by two investigators, one for each GP and TW3 methods in 85 normal children, 30 precocious puberty girls, and 30 constitutional growth delay boys. The differences between the measured predicted adult heights using the BP and TW3 methods were compared in each group. RESULTS: The bone age measured by the TW3 method was less than that by the GP method in normal children. The predicted adult heights measured by the two methods showed no significant difference in normal boys, while the predicted adult height measured by the TW3 method was higher than that by the BP method for normal girls (156.4+/-4.7 cm vs. 158.9+/-3.8 cm, P<0.01) and for precocious puberty girls (156.3+/-4.0 cm vs. 159.3+/-4.2 cm, P<0.01). In contrast, the predicted adult height was higher from the BP method than from the TW3 method in constitutional growth delay boys (173.3+/-4.4 cm vs. 169.7+/-3.2 cm, P<0.01). CONCLUSION: There were significant differences in predicted adult heights between the BP and TW3 method in normal girls, precocious puberty girls, or constitutional growth delay boys. In precocious puberty and constitutional growth delay, the BP method might be preferred to predict adult height, but further studies on final adult height are needed.