Associations between psychiatric disorders and mortality have been extensively studied, but limited evidence exists regarding influence of clinical characteristics on mortality risk, at the time of acute psychiatric hospitalization. A prospective total-cohort study included all patients consecutively admitted to Haukeland University Hospital's psychiatric acute ward in Bergen, Norway between 2005 and 2014 (n = 6125). Clinical interviews were conducted at the first admission within the study period, and patients were subsequently followed for up to 15 years in the Norwegian Cause of Death Registry. Competing risks regression models were used to investigate associations between clinical characteristics at first admission and the risk of natural and unnatural death during follow-up. The mean age at first admission and at time of death was 42.5 and 62.8 years, respectively, and the proportion of women in the sample was 47.2%. A total of 1381 deaths were registered during follow-up, of which 65.5% had natural, 30.4% unnatural, and 4.1% unknown causes. Higher age, male sex, unemployment, cognitive deficits, and physical illness were associated with increased risk of natural death. Male sex, having no partner, physical illness, suicide attempts, and excessive use of alcohol and illicit substances were associated with increased risk of unnatural death. Psychiatric symptoms, except suicide attempts, were unrelated to increased mortality risk. In the endeavor to reduce the increased mortality risk in people with mental disorders, focus should be on addressing modifiable risk factors linked to physical health and excessive use of alcohol and illicit substances.