Abstract Objective Firstly, to map out and compare all‐cause and cause‐specific mortality patterns by migrant background in Belgium; and secondly, to probe into explanations for the observed patterns, more specifically into the healthy‐migrant, acculturation and the migration‐as‐rapid‐health‐transition theories. Methods Data comprise individually linked Belgian census‐mortality follow‐up data for the period 2001–2011. All official inhabitants aged 25–54 at time of the census were included. To delve into the different explanations, differences in all‐cause and chronic‐ and infectious‐disease mortality were estimated using Poisson regression models, adjusted for age, socioeconomic position and urbanicity. Results First‐generation immigrants have lower all‐cause and chronic‐disease mortality than the host population. This mortality advantage wears off with length of stay and is more marked among non‐Western than Western first‐generation immigrants. For example, Western and non‐Western male immigrants residing 10 years or more in Belgium have a mortality rate ratio for cardiovascular disease of 0.72 (95% CI 0.66–0.78) and 0.59 (95% CI 0.53–0.66), respectively ( vs host population). The pattern of infectious‐disease mortality in migrants is slightly different, with rather high mortality rates in first‐generation sub‐Saharan Africans and rather low rates in all other immigrant groups. As for second‐generation immigrants, the picture is gloomier, with a mortality disadvantage that disappears after control for socioeconomic position. Conclusion Findings are largely consistent with the healthy‐migrant, acculturation and the migration‐as‐rapid‐health‐transition theories. The convergence of the mortality profile of second‐generation immigrants towards that of the host population with similar socioeconomic position indicates the need for policies simultaneously addressing different areas of deprivation.