Later sleep timing, circadian preference, and circadian rhythm timing predict worse outcomes across multiple domains, including mood disorders, substance use, impulse control, and cognitive function. Disturbed sleep is common among pregnant and postpartum women. We examined whether sleep timing during third trimester of pregnancy predicted postpartum symptoms of mania, depression, and obsessive-compulsive disorder (OCD). Fifty-one women with a previous, but not active, episode of unipolar or bipolar depression had symptoms evaluated and sleep recorded with wrist actigraphy at 33 weeks of gestation and 2, 6, and 16 weeks postpartum. Circadian phase was measured in a subset of women using salivary dim light melatonin onset (DLMO). We divided the sample into “early sleep” and “late sleep” groups using average sleep onset time at 33 weeks of gestation, defined by the median-split time of 11:27 p.m. The “late sleep” group reported significantly more manic and depressive symptoms at postpartum week 2. Longer phase angle between DLMO and sleep onset at 33 weeks was associated with more manic symptoms at postpartum week 2 and more obsessive-compulsive symptoms at week 6. Delayed sleep timing in this sample of at-risk women was associated with more symptoms of mania, depression, and OCD in the postpartum period. Sleep timing may be a modifiable risk factor for postpartum depression.