School-based water, sanitation and hygiene (WASH) may improve the health and attendance of schoolchildren, particularly post-menarcheal girls, but existing evidence is mixed. We examined the impact of an urban school-based WASH programme (Project WISE) on child health and attendance. The WISE cluster-randomised trial, conducted in 60 public primary schools in Addis Ababa, Ethiopia over one academic year, enrolled 2-4 randomly selected classes per school (~ 100 pupils) from grades 2 to 8 (aged 7-16) in an 'open cohort'. Schools were assigned 1:1 by stratified randomisation to receive the intervention during the 2021/2022 or the 2022/2023 academic year (waitlist control). The intervention included improvements to drinking water storage, filtration and access, handwashing stations and behaviour change promotion. Planned sanitation improvements were not realised. At four unannounced classroom visits post-intervention (March-June 2022), enumerators recorded primary outcomes of roll-call absence, and pupil-reported respiratory illness and diarrhoea in the past 7days among pupils present. Analysis was by intention-to-treat. Of 83 eligible schools, 60 were randomly selected and assigned. In total, 6229 eligible pupils were enrolled (median per school 101.5; IQR 94-112), 5987 enrolled at study initiation (23rd November-22nd December 2021) and the remaining 242 during follow-up. Data were available on roll-call absence for 6166 pupils (99.0%), and pupil-reported illness for 6145 pupils (98.6%). We observed a 16% relative reduction in odds of pupil-reported respiratory illness in the past 7days during follow-up in intervention vs. control schools (aOR 0.84; 95% CI 0.71-1.00; p = 0.046). There was no evidence of effect on pupil-reported diarrhoea in the past 7days (aOR 1.15; 95% CI 0.84-1.59; p = 0.39) nor roll-call absence (aOR 1.07; 95% 0.83-1.38; p = 0.59). There was a small increase in menstrual care self-efficacy (aMD 3.32 on 0-100 scale; 95% CI 0.05-6.59), and no evidence of effects on other secondary outcomes. This large-scale intervention to improve school WASH conditions city-wide had a borderline impact on pupil-reported respiratory illness but no effect on diarrhoeal disease nor pupil absence. Future research should establish relationships between WASH-related illness, absence and other educational outcomes. ClinicalTrials.gov, number NCT05024890.