Mid-upper arm circumference (MUAC) is used as an alternative measure for body mass index to determine thinness in older persons. However, there are limited data on the reproducibility of this measurement in an older population. The present study examined the reproducibility of MUAC measurements in older persons, as well as the influence of different body positions and clothing. A cross-sectional reproducibility study was performed in a nursing home (n = 43; age 65-96 years) and swimming pool facilities (n = 107; age 65-88 years). A different pair of observers independently measured the MUAC of each participant in the upright position on two occasions within 1 week. In the nursing home, measurements were also performed for each participant in the laying position and with clothes covering the upper arm. Mean differences and the 95% limit of agreement for inter-observer reproducibility of MUAC were 0.0 cm (-2.6 to 2.5 cm) for the swimming pool facilities and 0.3 cm (-0.6 to 1.3 cm) for the nursing home. Intra-class correlation coefficients (ICCs) were 0.89 and 0.92, respectively. Mean differences between laying and upright positions were 0.1 cm (-2.0 to 2.2 cm) and 0.0 cm (-1.9 to 2.0 cm) for each observer, respectively (ICC 0.96-0.97). Mean differences between clothes versus bare upper arm were -2.7 cm (-6.2 to 0.7) and -2.4 (-5.6 to 0.9 cm) (ICC 0.75 and 0.78). The reproducibility of the MUAC measurement in older persons is acceptable for group comparisons and, although borderline for the swimming pool facilities, remains acceptable for clinical purposes. The measurement can also be performed in the laying position but not with clothes covering the upper arm.