Objective: To identify the characteristics of and measures taken by Japanese medical facilities providing maternity services that have recorded cases of nosocomial infection with SARS-CoV-2, which causes coronavirus disease 2019 (COVID-19). Design: A nationwide questionnaire survey: case-control study. Setting: Medical facilities providing maternity service in Japan. Population: 1,418 medical facilities providing maternity service (65% of all delivery facilities in Japan). Methods: A nationwide questionnaire survey about nosocomial COVID-19 was conducted in July 2020. A case-control study was conducted by including medical facilities providing maternity service with (Cases) and without (Control) nosocomial COVID-19. Main outcome measures: Contributed characteristics and variables associated with nosocomial COVID-19. The number of treated pregnant women with COVID-19 until June 2020 and cases of nosocomial COVID-19 in Japan. Results: Seventy-five COVID-19-positive pregnant women were treated in 48 facilities. Responses from Nosocomial infection was reported in 4.1% of the facilities. Nine reports developed a nosocomial infection in the maternity ward or obstetric department. Variables that contributed to nosocomial COVID-19 were chest CT on admission (adjusted odds ratio [95% confidence interval], 4.76 [2.44–9.27]), PCR test for SARS-CoV-2 before delivery (2.27 [1.21–4.25]), transfer of pregnant women with fever to another hospital (0.21 [0.06–0.71]), and private clinics (0.17 [0.05–0.60]). Conclusion: Nosocomial infection is likely to occur in large hospitals that treat a higher number of patients than private clinics do. Nosocomial infection can occur even in medical facilities performing COVID-19 screening tests, such as chest CT and PCR test.