PURPOSE: Natural disaster has an impact on mental health. The 2018 Japan Floods, which took place in July 2018 were one of the largest water disasters in Japan's recorded history. We aimed to evaluate the change in the number of benzodiazepine prescriptions by physicians before and after the disaster. METHODS: A retrospective cohort study based on the National Database of Health Insurance Claims was conducted in the flood-stricken areas between July 2017 and June 2019. The subjects were divided between victims and non-victims according to certification by local governments. Members of both groups were then categorized into three groups based on their pre-flood use of benzodiazepines: non-user, occasional user, and continuous user. Difference-in-differences (DID) analysis with a logistic regression model was conducted to estimate the effect of the disaster among victims by comparing the occurrence of benzodiazepine prescriptions before and after the disaster. RESULTS: Of 5,000,129 people enrolled, 31,235 were victims. Among all participants, the mean prescription rate for benzodiazepines in victims before the disaster (11.3%) increased to 11.8% after the disaster, while that in non-victims (8.3%) decreased to 7.9%. The DID analysis revealed that benzodiazepine prescription among victims significantly increased immediately after the disaster (adjusted ratio of odds ratios (ROR) 1.07: 95% confidence interval 1.05-1.11), and the effect of the disaster persisted even 1 year after the disaster (adjusted ROR 1.2: 95% confidence interval 1.16-1.24). CONCLUSION: The flood increased the number of benzodiazepines prescriptions among victims, and the effect persisted for at least 1 year.