Postpartum depression (PPD) affects ~10-15% of childbearing individuals, with deleterious consequences for two generations. Recent research has explored the biological mechanisms of PPD, particularly neuroactive steroids (NAS). We sought here to investigate associations between NAS levels and ratios during pregnancy and the subsequent development of depressive symptoms with postpartum onset. NAS levels and psychological scales were measured in individuals with and without mood disorders at up to eight visits across pregnancy and postpartum. Generalized linear mixed-effects regression models were used to assess relationships in euthymic pregnant individuals between each of the NAS biomarkers and ratios and subsequent PPD. Participants with a one-unit increase in the log isoallopregnanolone/pregnanolone ratio at the third trimester (T3) had higher odds (OR = 1.64, 95% CI: 1.13-2.37, FDR adjusted p = 0.038, C-index = 0.82), and those with a one-unit increase in the log pregnanolone/progesterone ratio at T3 had lower odds (OR = 0.64, 95% CI: 0.47-0.88, FDR adjusted p = 0.036, C-index = 0.82) of developing PPD; those with a one-unit increase in the log progesterone level at T3 had higher odds of developing PPD (OR = 4.00, 95% CI: 1.54-10.37, FDR adjusted p = 0.035, C-index = 0.80). We found key differences in the progesterone metabolic pathway at the third trimester, indicating likely decreased activity/expression of the 3α-HSD enzyme and/or increased activity/expression of 3β-HSD.