Recent studies found positive associations between intake of red meat and processed meat and total mortality; however, substitution of red meat with poultry and fish has been poorly investigated. We aimed to investigate associations for substitutions of red meat (unprocessed/processed) and total mortality and deaths due to cancer or CVD. We used data from the Danish Diet, Cancer and Health cohort, including 57 053 participants aged 50–64 years at baseline. Information on diet was collected through a validated 192-item FFQ. Information regarding total mortality, deaths due to cancer and deaths due to CVD was obtained by record linkage. Cox proportional hazards models were used to estimate the hazard ratio (HR) of 150 g/week substitutions of red meat with poultry or fish. During a follow-up (mean 16·1 years), 8840 deaths occurred (4567 were due to cancer; 1816 due to CVD). The adjusted HR for total death when substituting 150 g/week total red meat with poultry was 0·96 (95 % CI 0·95, 1·00) and with fish 0·99 (95 % CI 0·97, 1·01). Corresponding HR for cancer death or CVD death were similar. Substitution of processed red meat with fish or poultry was more consistently associated with a lower mortality than substitution of unprocessed red meat. For example, the adjusted HR for total death when substituting 150 g/week processed red meat with poultry was 0·95 (95 % CI 0·92, 0·98). We found that replacing processed red meat with poultry or fish was associated with a lower risk of total mortality and deaths due to cancer, but not deaths due to CVD.