Background This study assessed the association between timing and course of maternal depression from pregnancy onwards and children's cognitive development at ages 5 to 6. Potential interaction effects with child sex and family socioeconomic status were explored. Methods One thousand thirty‐nine mother–child pairs from the French EDEN mother–child birth cohort were followed from 24 to 28 weeks of pregnancy onwards. Based on Center for Epidemiological Studies Depression (CES‐D) and Edinburgh Postnatal Depression Scale (EPDS) scores assessed at six timepoints, longitudinal maternal depressive symptom trajectories were calculated with a group‐based semiparametric method. Children's cognitive function was assessed at ages 5 to 6 by trained interviewers with the Wechsler Preschool and Primary Scale of Intelligence Third Edition (WPPSI‐III), resulting in three composite scores: Verbal IQ (VIQ), Performance IQ (PIQ), and Full‐Scale IQ (FSIQ). Results Five trajectories of maternal symptoms of depression could be distinguished: no symptoms, persistent intermediate‐level depressive symptoms, persistent high depressive symptoms, high symptoms in pregnancy only, and high symptoms in the child's preschool period only. Multiple linear regression analyses showed that, compared to children of mothers who were never depressed, children of mothers with persistent high levels of depressive symptoms had reduced VIQ, PIQ, and FSIQ scores. This association was moderated by the child's sex, boys appearing especially vulnerable in case of persistent maternal depression. Conclusions Chronicity of maternal depression predicts children's cognitive development at school entry age, particularly in boys. As maternal mental health is an early modifiable influence on child development, addressing the treatment needs of depressed mothers may help reduce the associated burden on the next generation.