This study explores the potential modifying effect of age and mediation effect of co-morbidity on the association between land use mix, a measure of neighbourhood walkability, and five-year mortality among the 2424 individuals participating in the year-10 follow-up of the Cognitive Function and Ageing Study in England. Postcodes of participants were mapped onto Lower-layer Super Output Areas, a small area level geographical unit in the UK, and linked to Generalised Land Use data. Cox regression models were fitted to investigate the association. For the younger older age group (75-79 years), the effect of high land use mix on an elevated risk of mortality was mediated by co-morbidity. For older old age groups (80-84, 85+ years), a higher land use mix was directly associated with a 10% lower risk of five-year mortality. The findings suggest differential impacts of land use mix on the health of the younger and older old.