Purpose: The purpose of this study was to examine the association between homelessness and psychiatric disorders, including substance use disorders, on one hand, and cause-specific and all-cause mortality on the other in a high-income country. Methods: A historical nationwide register-based cohort study of the Danish population from 15 years of age between 2000 and 2011 was conducted. The association between homelessness, psychiatric disorders, and mortality was analysed by Poisson Regression adjusting for important confounders. Standardised mortality ratios (SMRs) were calculated for people with a history of homelessness compared with the general population using direct age-standardisation. Results: During 51,892,324 person-years of observation, 656,448 died. People with at least one homeless shelter contact accounted for 173,592 person-years with 4345 deaths. The excess mortality in the population experiencing homelessness compared with the general population was reduced by 50% after adjusting for psychiatric diagnoses, including substance use disorders (mortality rate ratio (MRR) for men 3.30, 95% CI 3.18–3.41; women 4.41, 95% CI 4.14–4.71). Full adjustment including physical comorbidity and socioeconomic factors in a sub-cohort aged 15–29 years resulted in an MRR of 3.94 (95% CI 3.10–5.02) compared with the general population. The excess mortality associated with homelessness differed according to psychiatric diagnosis, sex, and cause of death. Conclusions: A mental health or a substance use disorder combined with homelessness considerably increases the risk of death. However, homelessness is also independently associated with high mortality. Differences in the association between homelessness and mortality in men and women suggest the need for sex-tailored interventions.