Purpose: This study was done to develop a prediction model for postpartum depression by verifying the mediation effect of antepartum depression. A hypothesized model was developed based on literature reviews and predictors of postpartum depression by Beck. Methods:Data were collected from 186 pregnant women who had a gestation period of more than 32 weeks and were patients at a maternity hospital,two obstetrics and gynecology specialized hospitals, or the outpatient clinic of K medical center. Data were analysed with descriptivestatistics, correlation and exploratory factor analysis using the SPSS/WIN 18.0 and AMOS 18.0 programs. Results: The final modifiedmodel had good fit indices. Parenting stress, antepartum depression and postpartum family support had statistically significant effects onpostpartum depression, and defined 74.7% of total explained variance of postpartum depression. Antepartum depression had significantmediation effects on postpartum depression from stress in pregnancy and self-esteem. Conclusion: The results of this study suggest thatit is important to develop nursing interventions including strategies to reduce parenting stress and improve postpartum family support inorder to prevent postpartum depression. Especially, it is necessary to detect and treat antepartum depression early to prevent postpartumdepression as antepartum depression can affect postpartum depression by mediating antepartum factors.