To examine the association between serum magnesium and incident heart failure (HF) in older men and investigate potential pathways including cardiac function, inflammation and lung function. Prospective study of 3523 men aged 60–79 years with no prevalent HF or myocardial infarction followed up for a mean period of 15 years, during which 268 incident HF cases were ascertained. Serum magnesium was inversely associated with many CVD risk factors including prevalent atrial fibrillation, lung function (FEV1) and markers of inflammation (IL-6), endothelial dysfunction (vWF) and cardiac dysfunction [NT-proBNP and cardiac troponin T (cTnT)]. Serum magnesium was inversely related to risk of incident HF after adjustment for conventional CVD risk factors and incident MI. The adjusted hazard ratios (HRs) for HF in the 5 quintiles of magnesium groups were 1.00, 0.72 (0.50, 1.05), 0.85 (0.59, 1.26), 0.76 (0.52, 1.11) and 0.56 (0.36, 0.86) respectively [p (trend) = 0.04]. Further adjustment for atrial fibrillation, IL-6, vWF and FEV1 attenuated the association but risk remained significantly reduced in the top quintile (≥ 0.87 mmol/l) compared with the lowest quintile [HR 0.62 (0.40, 0.97)]. Adjustment for NT-proBNP and cTnT attenuated the association further [HR 0.70 (0.44, 1.10)]. The benefit of high serum magnesium on HF risk was most evident in men with ECG evidence of ischaemia [HR 0.29 (0.13, 0.68)]. The potential beneficial effect of high serum magnesium was partially explained by its favourable association with CVD risk factors. Further studies are needed to investigate whether serum magnesium supplementation in older adults may protect from the development of HF.