Abstract Objective: To quantify the relationship between screening positive for depression and several indicators of the food and nutrition environment in Bangladesh. Design: We used cross-sectional data from the Food and Agricultural Approaches to Reducing Malnutrition (FAARM) trial in Bangladesh to examine the association of depression in non-peripartum (NPW) and peripartum women (PW) with food and nutrition security using multivariable logistic regression and dominance analysis. Setting: Rural north-eastern Bangladesh. Participants: Women of reproductive age. Results: Of 2599 women, 40 % were pregnant or up to 1 year postpartum, while 60 % were not peripartum. Overall, 20 % of women screened positive for major depression. In the dominance analysis, indicators of food and nutrition security were among the strongest explanatory factors of depression. Food insecurity (HFIAS) and poor household food consumption (FCS) were associated with more than double the odds of depression (HFIAS: NPW OR = 2·74 and PW OR = 3·22; FCS: NPW OR = 2·38 and PW OR = 2·44). Low dietary diversity (<5 food groups) was associated with approximately double the odds of depression in NPW (OR = 1·80) and PW (OR = 1·99). Consumption of dairy, eggs, fish, vitamin A-rich and vitamin C-rich foods was associated with reduced odds of depression. Anaemia was not associated with depression. Low BMI (<18·5 kg/m2) was also associated with depression (NPW: OR = 1·40). Conclusions: Depression among women in Bangladesh was associated with many aspects of food and nutrition security, also after controlling for socio-economic factors. Further investigation into the direction of causality and interventions to improve diets and reduce depression among women in low- and middle-income countries are urgently needed.