BACKGROUND: Maternal depression is associated with poor child development and growth in low-income countries. This paper evaluates the effect of a community-based trial providing psychosocial stimulation and food supplements to severely malnourished children on maternal depressive symptoms in Bangladesh. METHODS: Severely underweight (weight-for-age Z-score < -3) hospitalized children aged 6-24 months (n = 507), were randomly assigned to: psychosocial stimulation (PS), food supplementation (FS), PS+FS, clinic control (CC) and hospital control (CH) at discharge. PS included play sessions with children and parental counselling to mothers during fortnightly follow-up visit at community clinics, conducted by trained play leaders for 6 months. FS involved cereal-based supplements (150-300 kcal/day) for 3 months. All groups received medical care, micronutrient supplements and growth monitoring. We used Bayley scales, Home Observation for Measurement of Environment (HOME) inventory and a parenting questionnaire to assess child development, home stimulation and mothers' child-rearing practices, respectively. We assessed mothers' depressive symptoms using a modified version of Centre for Epidemiologic Studies Depression Scale at baseline and at 6 months post intervention. RESULTS: Maternal depressive symptoms were significantly lower in the CH group at baseline (P = 0.014). After 6 months of intervention there was no significant effect of intervention after adjusting for baseline scores and all possible confounders. Maternal depressive symptoms were higher among poorer (P = 0.06), older (P = 0.057) and less educated (P = 0.019) mothers, who were housewives (P = 0.053), and whose husbands had more unstable jobs (P = 0.058). At 6 months post intervention, children's cognitive (P = 0.045) and motor (P = 0.075) development, HOME (P = 0.012) and mother's parenting score (P = 0.057) were higher among mothers with lower depressive symptoms. CONCLUSION: The study did not show a significant effect of the intervention on the level of maternal depressive symptoms. Interventions with higher intensity and/or of longer duration focusing directly on maternal psychosocial functioning are probably needed to reduce maternal depressive symptoms.