BackgroundResearch using objectively measured physical activity (PA) in older adults to assess relationship between PA levels and mortality is scarce.ObjectiveTo investigate associations between level of physical activity and mortality in older adults over a 4-year period. MethodsThe population-based cohort study was carried out including 554 older participants (mean age: 76.2 ± 8.05 years) using data from the SABE study (Health, well-being and aging). Levels of physical activity were measured using accelerometers and participants were divided into tertiles and then categorized into two groups: (I) low level of physical activity and (II) intermediate/high level of physical activity. The dependent variable was mortality between 2010 and 2014. Control variables included socio-demographic and clinical factors. Multiple regression analysis was used from a hierarchical model, grouping the variables into two blocks ordered according to the magnitude of their effect. ResultsOur results showed that mortality rate in participants with low level of physical activity was 20/1000 person/year and for those with intermediate/high levels of physical activity was 14/1000 person/year. In the adjusted model, by sociodemographic and clinical variables, those with low levels of physical activity presented a higher risk for mortality (OR = 2.79, 95%CI = 1.71–4.57) when compared to individuals with intermediate/high levels of physical activity. ConclusionOlder adults with low levels of physical activity have a higher chance of mortality as compared to those with intermediate/high levels of physical activity, regardless of sociodemographic and clinical variables.