Background: We investigated the association between 10 years of change in physical activity (PA) levels and 7-year all-cause mortality. Methods: Mexican American adults aged 67 and older (N = 803) participating in the Hispanic Established Population for the Epidemiologic Study of the Elderly (1995/1996–2012/2013) were included. All-cause mortality was our outcome. Participants were divided into 4 groups based on their difference in overall PA between 1995/1996 and 2005/2006 measured by the Physical Activity Scale for the Elderly. Participants were classified as having unchanged low PA (n = 339), decreased activity (n = 233), unchanged high activity (n = 166), and increased activity (n = 65). Change in the frequency of PA domains was also investigated. PA domains included leisure, household, sedentary, and walking activities. Results:After controlling for all covariates, results from the Cox proportional hazards regression found a 43% lower mortality risk in the increased PA group (hazards ratio = 0.57; 95% confidence interval, 0.34–0.97) compared with the unchanged low PA group. In the entire sample, a significantly lower mortality risk was also present among walking (hazards ratio = 0.88) and household (hazards ratio = 0.88) activities. Conclusion:Our results suggest that, independent of other factors, increasing PA is most protective of mortality among older Mexican Americans.