While factors affecting smoking are well documented, the role of religion has received little attention. This national study aims to assess the extent to which religious affiliation is associated with current-smoking and ever-smoking, controlling for age, sex, ethnicity and socio-economic status. Variations between adult and youth populations are examined using secondary analysis of individual-level data from 5 years of the Health Survey for England for adult (aged >20, n = 39,837) and youth (aged 16–20, n = 2355) samples. Crude prevalence statistics are contrasted with binary logistic models for current-smoking and ever-smoking in the adult and youth samples. Analyses suggest that Muslims smoke substantially less than Christians. Highest levels of smoking characterise people not professing any religion. Associations between smoking and the Muslim religion attenuate to statistical insignificance in the face of ethnic and socio-economic factors. An association between smoking and the absence of a religious affiliation is sustained. An understanding of the association between smoking and religion is essential to the development of tobacco control programmes.