Aim: Although the Edinburgh Postnatal Depression Scale (EPDS) is widely used for screening for postpartum depression (PPD), screening time point and cut-off value in the current medical environment are not sufficiently examined. In this study, we examined appropriate timing and cut-off value of EPDS for the diagnostic criteria of Diagnostic and Statistical Manual of Mental Disorders Fifth Edition. Methods: This study was a prospective cohort study of women who delivered at the Iwate Medical University Hospital from August 2016 to February 2019. The EPDS self-assessment was conducted at 4 days and at 1 month after delivery (4D-EPDS and 1M-EPDS), and the Center for Epidemiologic Studies-Depression scale (CES-D) self-assessment was conducted at 1 month after delivery (1M-CES-D) in eligible patients and the scores were tabulated. Participants were interviewed by a psychiatric specialist to identify PPD within 2 months after delivery. For each screening, a receiver operating characteristic analysis was performed to calculate the cut-off value for PPD. Sensitivity, specificity, positive predictive value, negative predictive value and correct diagnostic rate were evaluated. Results: Of the 80 post-partum women, PPD was found in nine cases (11.3%). The cut-off values for PPD were ≥10 for 4D-EPDS, ≥13 for 1M-EPDS and ≥21 for 1M-CES-D. No significant difference was found in the sensitivity, positive predictive value or negative predictive value among the three screening methods. However, a significant difference was found in the specificity, and the correct diagnosis rate and these values were relatively higher in 1M-EPDS. Conclusion: Based on the result of this Japanese cohort, 1M-EPDS with a cut-off value of ≥13 might be appropriate for PPD screening although that of ≥9 is currently employed.