Objective: To compare the impact on child diet and growth of a multisectoral community intervention v. nutrition education and livestock management training alone., Design: Longitudinal community-based randomized trial involving three groups of villages assigned to receive: (i) Full Package community development activities, delivered via women's groups; (ii) livestock training and nutrition education alone (Partial Package); or (iii) no intervention (Control). Household surveys, child growth monitoring, child and household diet quality measures (diet diversity (DD), animal-source food (ASF) consumption) were collected at five visits over 36 months. Mixed-effect linear regression and Poisson models used survey round, treatment group and group-by-round interaction to predict outcomes of interest, adjusted for household- and child-specific characteristics., Setting: Banke, Nepal., Participants: Households (n 974) with children aged 1-60 months (n 1333)., Results: Children in Full Package households had better endline anthropometry (weight-for-age, weight-for-height, mid-upper-arm-circumference Z-scores), DD, and more consumption of ASF, after adjusting for household- and child-specific characteristics. By endline, compared with Partial Package or Control groups, Full Package households demonstrated preferential child feeding practices and had significantly more improvement in household wealth and hygiene habits., Conclusions: In this longitudinal study, a comprehensive multisectoral intervention was more successful in improving key growth indicators as well as diet quality in young children. Provision of training in livestock management and nutrition education alone had limited effect on these outcomes. Although more time-consuming and costly to administer, incorporating nutrition training with community social capital development was associated with better child growth and nutrition outcomes than isolated training programmes alone.