This study aimed to investigate the association between maternal depressive symptoms and anxiety and interference in the mother/child relationship, using structural equations modeling. Data were used from a prospective cohort study initiated during the prenatal period with 1,140 mothers in São Luís, Maranhão State, Brazil. Data were collected during prenatal care and when the children reached two years of age. Interference in the mother/child relationship was measured with the Postpartum Bonding Questionnaire - PBQ (N = 1,140). In the initial theoretical model, socioeconomic status determined the maternal demographic, psychosocial, and social support factors, which determined the outcome, i.e., the mother/child relationship. Adjustments were performed by structural equations modeling, using Mplus 7.0. The final model showed good fit (RMSEA = 0.047; CFI = 0.984; TLI = 0.981). Depressive symptoms in pregnancy and the postpartum were associated with higher PBQ scores, indicating interference in the mother/child relationship. The greatest effect was from depressive symptoms in pregnancy. Other factors associated with higher PBQ scores were lower social support, unfavorable socioeconomic status, and living without a partner, by indirect association. Anxiety symptoms and maternal age were not associated with the mother/child relationship. The results suggest that identifying and treating depression in pregnancy and postpartum can improve mother/child bonding in childhood.