We examined the relationship between maternal perinatal depression trajectories and infant development considering various ecological risk factors in the environment and employing growth mixture modelling. Participants were 717 mother-child dyads in rural Malawi. The mothers were interviewed three times: during pregnancy, 1-year postpartum, and 2-year postpartum. Their children were administered the Bayley test. First, because of the latent class growth model, a two-class solution was adopted: a 'depression mitigated group' (90.2%) and a 'depression aggravated group' (9.8%). Second, mothers' depression intensified and was maintained by risk factors that were not only intrapersonal but also contextual. Finally, infants in the 'depression aggravated group' scored significantly lower in cognitive and motor development than did infants in the 'depression mitigated group', suggesting the need for various support programmes including psychosocial interventions (e.g. women's empowerment and parenting education) beyond merely providing individual, basic treatment to support the healthy development of mothers and children.