This review summarises evidence for an association between vitamin D status and CVD and the mechanisms involved. Vitamin D3 is predominantly provided by the action of UVB from sunlight on skin. Average UK diets supply 2–3 μg/d vitamin D but diets containing at least one portion of oily fish per week supply about 7 μg/d. Pharmacological doses of vitamin D2 (bolus injection of 7500 μg or intakes >50 μg/d) result in a smaller increase in plasma 25(OH)D than those of D3 but physiological doses 5–25 μg/d seem equivalent. Plasma 25(OH)D concentrations are also influenced by clothing, obesity and skin pigmentation. Up to 40 % of the population have plasma 25(OH)D concentrations <25 nmol/l in the winter compared with <10 % in the summer. The relative risk of CVD death is 1·41 (95 % CI 1·18, 1·68) greater in the lowest quintile of plasma 25(OH)D according to meta-analysis of prospective cohort studies. Acute deficiency may inhibit insulin secretion and promote inflammation thus increasing the risk of plaque rupture and arterial thrombosis. Chronic insufficiency may increase arterial stiffness. There is no evidence to support claims of reduced CVD from existing trials with bone-related health outcomes where vitamin D was usually co-administered with calcium. Although several trials with cardiovascular endpoints are in progress, these are using pharmacological doses. In view of the potential toxicity of pharmacological doses, there remains a need for long-term trials of physiological doses of D2 and D3 with CVD incidence as the primary outcome.