Arm anthropometry is used as a proxy of body composition in clinical and field research but its validity has not been established in children. To address this issue, mid-upper arm circumference (MUAC) and triceps skinfold thickness (TS) were measured in 110 healthy children aged 4.4-13.9 y (55 boys) and 49 cystic fibrosis (CF) patients aged 8.1-13.4 y (22 boys). Reference values were arm and whole-body fat mass (FM) and fat-free mass (FFM) measured by dual x-ray absorptiometry and four-component model, respectively. Arm fat area (AFA), MUAC, and TS correlated well with arm FM (r = 0.84-0.92) and total FM (r = 0.78-0.92). Arm muscle area (AMA) and MUAC correlated well with arm FFM (r = 0.68-0.82) and total FFM (r = 0.60-0.86). After adjusting for age, sex, and height, arm anthropometry correlated strongly with FM but weakly with FFM. AFA, MUAC, and TS explained 67, 63, and 61% of variability in total FM in healthy children and 70, 72, and 63% in CF. AMA and MUAC explained only 24 and 16% of variability in total FFM in healthy children and 33 and 28% in CF. Arm anthropometry is useful for predicting FM and ranking healthy children and patients for fatness. It has poorer success in predicting regional or total FFM.