Background: Interventions to improve early childhood development have previously addressed only one or a few risk factors. Learning Clubs is a structured, facilitated, multicomponent programme designed to address eight potentially modifiable risk factors, and offered from mid-pregnancy to 12 months post partum; we aimed to establish whether this programme could improve the cognitive development of children at 2 years of age. Methods: For this parallel-group cluster-randomised controlled trial, 84 of 116 communes (the clustering unit) in HaNam Province in rural Vietnam were randomly selected and randomly assigned to receive the Learning Clubs intervention (n=42) or usual care (n=42). Women aged at least 18 years who were pregnant (gestational age <20 weeks) were eligible for inclusion. Data sources were standardised, and study-specific questionnaires assessing risks and outcomes were completed in interviews in mid-pregnancy (baseline), late pregnancy (after 32 weeks of gestation), at 6–12 months post partum, and at the end of the study period when children were 2 years of age. Mixed-effects models were used to estimate trial effects, adjusting for clustering. The primary outcome was the cognitive development of children at 2 years of age, assessed by the Bayley Scales of Infant and Toddler Development Third Edition (Bayley-III) cognitive score. This trial is registered with the Australian New Zealand Clinical Trials Registry (ACTRN12617000442303). Findings: Between April 28, 2018, and May 30, 2018, 1380 women were screened and 1245 were randomly assigned (669 to the intervention group and 576 to the control group). Data collection was completed on Jan 17, 2021. Data at the end of the study period were contributed by 616 (92%) of 669 women and their children in the intervention group, and by 544 (94%) of 576 women and their children in the control group. Children aged 2 years in the intervention group had significantly higher mean Bayley-III cognitive scores than those in the control group (99·6 [SD 9·7] vs 95·6 [9·4]; mean difference 4·00 [95% CI 2·56–5·43]; p<0·0001). At 2 years of age, 19 (3%) children in the intervention group had Bayley-III scores less than 1 SD, compared with 32 (6%) children in the control group, but this difference was not significant (odds ratio 0·55 [95% CI 0·26–1·17]; p=0·12). There were no significant differences between groups in maternal, fetal, newborn, or child deaths. Interpretation: A facilitated, structured, community-based, multicomponent group programme improved early childhood development to the standardised mean in rural Vietnam and could be implemented in other similarly resource-constrained settings. Funding: Australian National Health and Medical Research Council and Grand Challenges Canada Saving Brains Initiative. Translation: For the Vietnamese translation of the abstract see Supplementary Materials section.