Setting: Seven pilot sites in Zimbabwe implementing 6 months of isoniazid preventive therapy (IPT) for people living with the human immunodeficiency virus (PLHIV). Objectives: To determine, among PLHIV started on IPT, the completion rates for a 6-month course of IPT and factors associated with non-adherence. Design: A retrospective cohort study. Results: Of 578 patients, 466 (81%) completed IPT. Of the 112 patients who failed to complete IPT, 69 (60%) were lost to follow-up, 30 (27%) stopped treatment with no documented reasons, 8 (7%) developed toxicity/adverse reactions, 5 (5%) were documented as having drug stock-outs and the remainder transferred out or refused to continue treatment. Currently being on antiretroviral therapy (ART) (aOR 0.09, 95%CI 0.03-0.28) and receiving a ⩾2 month supply of isoniazid at the start of treatment were associated with a lower risk of not completing IPT, while missing clinic visits prior to starting IPT (aOR 5.25, 95%CI 2.10-13.14) was associated with a higher risk of non-completion. Conclusion: IPT completion rates in seven pilot sites of Zimbabwe were comparatively high, showing that IPT roll-out in public health facilities is feasible. Enhanced adherence counselling or active tracing among pre-ART patients and those with a history of loss to follow-up may improve IPT completion rates, along with synchronising IPT and ART resupplies.