The aim of the study was to determine the sensitivity and specificity of the BMI cut-off points that have been applied in the National Nutritional Survey in Japan (NNS-J) for young Japanese females. The relationship between the body mass index (BMI) and percentage body fat (%BF) of 139 young Japanese females (aged 18-27 years old) was examined using dual energy x-ray absorptiometry (DXA) and anthropometry. The sensitivity and specificity of the BMI cut-off point that has been in use in Japan was compared with the cut-off point proposed for Asia-Pacific population (the WPRO classification). The regression equation calculated for the relationship between the BMI and %BF using the DXA scan data was %BF = -15.152+2.058*BMI; R2 = 0.612, SEE = 3.66. From this equation the BMI value that corresponded with %BF of 30% was calculated to be 21.9kg/m2 (95%CI: 16.9-28.6). From the sensitivity and specificity analysis, it was found that more than 90% of Japanese females with %BF >30% were misclassified as "not obese" when the BMI value of 25kg/m2 was used to identify obese individuals. The misclassification reduced to 60% when the BMI value of 23kg/m2 was used as a cut-off point. The results indicate that the current classification used in the NNS-J may not precisely reflect the obesity status of young Japanese females and a considerable proportion of females with a large fat deposition would be misclassified as not obese. The current study shows the importance of including detailed body composition assessments to determine obesity level of individuals.