Low vitamin D associated with high parathyroid hormone (PTH) is common in HIV infection. We determined the association between total 25(OH)D and PTH in adolescents living with HIV, in Zambia and Zimbabwe. Adolescents (11–19 years) perinatally infected with HIV and established on antiretroviral therapy for ≥ 6 months were recruited into a cross-sectional study. Socio-demographic and clinical characteristics were recorded, anthropometry measured and fasted serum concentrations of 1,25(OH)2D, total 25(OH)D and intact PTH measured. The association between total 25(OH)D and PTH was examined using natural cubic spline regression. 842 participants (female: 53·2%) with a median age of 15·5 (IQR: 13·2–17·9) years were enrolled. Median antiretroviral therapy duration was 9·8 (IQR: 6·3–12·3) years, and 165/841 had an HIV viral-load >60 copies/ml. Stunting (height-for-age z-score <–2) and underweight (weight-for-age z-score <–2) were observed in 29·9 and 30·0%, respectively. Three-quarters reported daily Ca intakes <150 mg/d. The mean (sd) concentrations of total 25(OH)D and 1,25(OH)2D were 66·1(16·5) nmol/l and 210·6 (70·4) pmol/l, respectively, and median PTH level was 4·3 (IQR: 3·3–5·5) pmol/l. There was an inverse non-linear relationship between total 25(OH)D and PTH, 25(OH)D levelling off at 74·6 nmol/l (95 % CI: 74·5, 75·2). Results were consistent in those taking tenofovir disoproxil fumarate and virally unsuppressed participants. In this population with extremely low habitual Ca intakes, the lack of association between 25(OH)D and PTH when 25(OH)D exceeded 75 nmol/l potentially suggests that levels of 25(OH)D >75 nmol/l may need to be achieved to improve bone health; investigation is needed in future research studies.