Background: The Edinburgh Postnatal Depression Scale (EPDS) has been validated in many countries, but not in Mainland China. Objectives: This study investigated the reliability and validity of the Mainland Chinese version of the EPDS. Design: A three-stage design was used for this study. Stage I consisted of a multi-stepped process of forward and backward translation, using a panel of six experts to test content validity. Stage II established the psychometric properties of the EPDS by examining the convergent, discriminant and construct validity, internal consistency and stability of the scale. Stage III established its sensitivity, specificity and the optimal cutoff score of the EPDS according the DSM-IV-TR criteria using the Structured Clinical Interview. Setting: Three regional public hospitals in Chengdu. Participants: A convenience sample was composed of 312 and 451 pregnant women at 28-36 weeks' gestation. Method: Translation and back-translation of the original English instrument and content validation by an expert panel. Receiver operating characteristic (ROC) analysis was carried out to evaluate the global functioning of the scale. The Beck Depression Inventory (BDI), Dyadic Adjustment Scale (DAS) and standard SF-12 Health Survey (SF-12) were used to investigate the convergent, discriminant and construct validity. The Cronbach's alpha reliability coefficient, split-half reliability and test-retest reliability were used to examine the internal consistency and stability of the scale. Results: The translation process was rigorously conducted to ensure that equivalence was established. Content validity was confirmed by a satisfactory level of agreement with a content validity index (CVI) of 0.93. The area under curve (AUC) was 85.6% and the logistic estimate for the threshold score was 9.5 (sensitivity, 80.0%; specificity, 83.03%) for clinical depression. Convergent and construct validity was supported and discriminant validity suggested that the EPDS successfully discriminated among the non-depressed, mildly and clinically depressed groups. The split-half reliability of the EPDS was 0.76, Cronbach's alpha was 0.79 and test-retest reliability was 0.85. Conclusions: Based on the results of this psychometric testing, the Mainland Chinese version of the EPDS is considered ready for use in the routine screening of pregnant women. It is hoped that with this type of cross-cultural information, the universality of the constructs of the instrument can be reliably demonstrated. © 2009 Elsevier Ltd. All rights reserved.