BACKGROUND Adiposity may contribute to the future risk of disease. The aim of this study was to evaluate the accuracy and reliability of an air-displacement plethysmography (ADP) system to estimate percentage fat mass (%FM) in preterm infants and to evaluate interdevice reliability in infants. METHODS A total of 70 preterm and 9 full-term infants were assessed. The accuracy of ADP measurements was assessed by determining reference %FM values using H(2)(18)O dilution measurement. RESULTS Mean %FM by ADP was 5.67 ± 1.84 and mean %FM by H(2)18O dilution was 5.99 ± 2.56. Regression analysis showed that %FM by ADP was associated with %FM by H(2)(18)O dilution (R2 = 0.63, SE of estimate (SEE) = 1.65, P = 0.006). Bland-Altman analysis showed no bias (r = -0.48, P = 0.16) and 95% limits of agreement were -3.40 to 2.76 %FM. There was no difference in mean interdevice reliability %FM values (8.97 vs. 8.55 %FM) between ADP 1 and 2. Regression analysis indicated a low SEE (1.14% FM) and high R2 (0.91); 95% limits of agreement were -1.87 to 2.71 %FM. The regression line did not differ significantly from the line of identity. CONCLUSIONS ADP is a noninvasive, reliable, and accurate technique to measure preterm infants' body composition in both research and clinical settings.