The current study examined factors associated with the increased risk of self-interruption of medications for mental disorders in pregnancy. We reviewed the obstetric records of all singleton deliveries beyond 22 weeks of gestation at Japanese Red Cross Katsushika Maternity Hospital form January 2014 through July 2018. We examined the clinical and social characteristics and obstetric outcomes in women with self-interruption of medications for mental disorders in pregnancy in comparison with those in women who continued the medications throughout the pregnancy. There were 177 women who were diagnosed with mental disorders requiring medications before pregnancy by Japanese psychiatric specialists. Of these, 46 (26%) self-interruptedi their medications during the first trimester of their pregnancies. The rates of partners with mental disorders (odds ratio: 4.39, p = 0.01) and presence of social support (odds ratio: 2.50, p = 0.02) in women showing self-interruption of their medications were significantly higher than those in women continuing the medications. In the presence of both factors of a partner's mental disorders and social support, the odds ratio for self-interruption of medications increased to 15.9 (95% confidence interval 2.4-100, p < 0.01). We believe that it may be the possible to identify women at high risk of self-interruption of medications during pregnancy. In women requiring perinatal mental health care, mental health support of their partners may also be needed.