The debate regarding the need for hospital evacuation and the evacuation distance remains rather chaotic. Furthermore, the relationship between hospital evacuation and the prognoses of psychiatric inpatients has not yet been investigated. We aimed to reveal the association between the long-term prognosis of psychiatric inpatients evacuated immediately following the Fukushima Daiichi Nuclear Power Plant accident and their backgrounds. In this retrospective cohort study, 777 psychiatric inpatients who were immediately evacuated from their hospitals following the accident were included for analysis. Survival time was the primary outcome. We conducted univariable and multivariable analyses to examine the associations between mortality and linear distance of evacuation and different backgrounds, including psychiatric/physical traits. Univariable analysis showed that the estimated survival time among patients was significantly associated with their evacuation distance. A multivariable analysis showed that a longer evacuation distance had a significantly lower hazard ratio (HR) and resulted in lower mortality. In contrast, older patients with physical complications of respiratory disease (International Statistical Classification of Diseases and Related Health Problems 10th revision, J00–99) and genitourinary disease (N00–99) showed a significantly higher HR and had a higher mortality than patients without these complications. To prevent death among elderly psychiatric inpatients with physical comorbidities during disasters, the evacuation destination should be determined taking into consideration the evacuees’ tolerance for long-distance transportation and the availability of post-evacuation care in the destination hospitals.