BACKGROUND: HIV testing is the important entry point for HIV care and prevention service, but uptake of HIV testing and thus coverage of antiretroviral therapy are much lower in older children and adolescents than in adults. We investigated the effect of economic incentives provided to caregivers of children aged 8-17 years on uptake of HIV testing and counselling in Harare, Zimbabwe. METHODS: This randomised controlled trial was nested within a household HIV prevalence survey of children aged 8-17 years in Harare. Households with one or more survey participants whose HIV status was unknown were eligible to participate in the trial. Eligible households were randomly assigned (1:1:1) to either receive no incentive, receive a fixed US$2 incentive, or participate in a lottery for $5 or $10 if the participant presented for HIV testing and counselling at a local primary health-care centre. The survey fieldworkers who enrolled participants were not blinded to trial arm allocation, but the statistician was blinded for analysis of outcome. The primary outcome was the proportion of households in which at least one child had an HIV test within 4 weeks of enrolment. HIV test uptake in the incentivised groups was compared with uptake in the non-incentivised group using logistic regression, adjusting for community and number of children as fixed effects and research assistant as a random effect. All analyses were by intention to treat. The trial is registered with the Pan African Clinical Trials Registry, number PACTR201605001615280. FINDINGS: Between Aug 4, and Dec 18, 2015, 2050 eligible households were enrolled in the prevalence survey. 649 (32%) households were assigned no incentive, 740 (34%) households were assigned a $2 incentive, and 661 (32%) households were assigned to lottery participation. Children were unavailable in 148 households in the no-incentive group, 63 households in the $2 incentive group, and 81 households in the lottery group. 1688 households had at least one child with unknown HIV status and were enrolled into the trial. 22 households had no undiagnosed child, and one household refused consent. The primary outcome of HIV testing was assessed in 472 (28%) households in the no-incentive group, 654 (39%) households in the $2 incentive group, and 562 (33%) households in the lottery group. At least one child was HIV tested in 93 (20%) households in the no-incentive group, in 316 (48%) households in the $2 incentive group (adjusted odds ratio 3·67, 95% CI 2·77-4·85; p<0·0001), and in 223 (40%) of 562 households in the lottery group (2·66, 2·00-3·55; p<0·0001). No adverse events were reported. INTERPRETATION: Fixed incentives and lottery-based incentives increased the uptake of HIV testing by older children and adolescents, a key hard-to-reach population. This strategy would be sustainable in the context of vertical HIV infection as repeated testing would not be necessary until sexual debut. FUNDING: Wellcome Trust.