Background: The effect of prenatal secondhand smoke (SHS) exposure on childhood atopic dermatitis (AD) remains controversial. We aimed to investigate the association between prenatal SHS and childhood AD in a general population-based birth cohort. Methods: Patients included 2,360 mother–child pairs from the Cohort for Childhood Origin of Asthma and Allergic diseases (COCOA), stratified into 0–3, 4–6, and 7–9 years age groups. Prenatal SHS exposure was assessed using questionnaires. AD diagnosis and symptom assessments were conducted through annual visits by pediatric allergists. Skin prick tests for 18 allergens were conducted. Serum total IgE and eosinophil levels were measured at birth and ages 3 and 7 years. Maternal urine cotinine concentrations were measured at week 36 of gestation. Multivariate logistic regression was performed. Results: Children aged 7–9 years exposed to prenatal SHS were significantly more likely to have an AD diagnosis (aOR 1.670, 95% CI: 0.995–2.804) and current AD (aOR 1.823, 95% CI: 1.051–3.161). This association in AD diagnosis was stronger in children with sensitization (aOR 2.205, 95% CI: 1.048–4.642). Higher maternal urine cotinine levels increased the risk of current AD at ages 4–6 (aOR 2.816, 95% CI: 1.053–7.529). Children exposed to prenatal SHS were more likely to have a late-onset phenotype of AD (aOR 1.663, 95% CI: 1.038–2.664). Conclusion: SHS exposure during pregnancy was associated with late childhood AD. Prevention of prenatal SHS exposure is necessary to reduce the risk of AD in schoolchildren.