BackgroundMaternal and child undernutrition have adverse consequences for pregnancy outcomes and child morbidity and mortality, and they are associated with low educational attainment, economic productivity as an adult, and human wellbeing. ‘Nutrition-sensitive’ agriculture programs could tackle the underlying causes of undernutrition.Methods/designThis study is a four-arm cluster randomised controlled trial in Odisha, India. Interventions are as follows: (1) an agricultural extension platform of women’s groups viewing and discussing videos on nutrition-sensitive agriculture (NSA) practices, and follow-up visits to women at home to encourage the adoption of new practices shown in the videos; (2) women’s groups viewing and discussing videos on NSA and nutrition-specific practices, with follow-up visits; and (3) women’s groups viewing and discussing videos on NSA and nutrition-specific practices combined with a cycle of Participatory Learning and Action meetings, with follow-up visits. All arms, including the control, receive basic nutrition training from government community frontline workers. Primary outcomes, assessed at baseline and 32 months after the start of the interventions, are (1) percentage of children aged 6–23 months consuming ≥ 4 out of 7 food groups per day and (2) mean body mass index (BMI) (kg/m2) of non-pregnant, non-postpartum (gave birth > 42 days ago) mothers or female primary caregivers of children aged 0–23 months. Secondary outcomes are percentage of mothers consuming ≥ 5 out of 10 food groups per day and percentage of children’s weight-for-height z-score < -2 standard deviations (SD).The unit of randomisation is a cluster, defined as one or more villages with a combined minimum population of 800 residents. There are 37 clusters per arm, and outcomes will be assessed in an average of 32 eligible households per cluster. For randomisation, clusters are stratified by distance to nearest town (< 10 km or ≥ 10 km), and low (< 30%), medium (30–70%), or high (> 70%) proportion of Scheduled Tribe or Scheduled Caste (disadvantaged) households. A process evaluation will assess the quality of implementation and mechanisms behind the intervention effects. A cost-consequence analysis will compare incremental costs and outcomes of the interventions.DiscussionThis trial will contribute evidence on the impacts of NSA extension through participatory, low-cost, video-based approaches on maternal and child nutrition and on whether integration with nutrition-specific goals and enhanced participatory approaches can increase these impacts.Trial registrationISRCTN , ISRCTN65922679. Registered on 21 December 2016.Electronic supplementary materialThe online version of this article (10.1186/s13063-018-2521-y) contains supplementary material, which is available to authorized users.