The purpose of this study was to examine relationships between participation in physical activities and long-term survival in older men and women. Cox proportional hazards regression models were used to investigate participation in different physical activities on 21-year all-cause mortality risk. The NLSAA survey was conducted in 1985 with 1,042 participants. As of 31st May 2006 (21 years from baseline), there were 919 recorded deaths (368 men and 551 women). Mortality analyses were from baseline interview to death or censorship (May 2006). Men and women were analysed separately in unadjusted models and adjusting for demographic and health-related variables. In separate unadjusted models, participation in outdoor activities, indoor productive activities, leisure activities, walking, undertaking activities requiring flexibility and physical effort each were associated with reduced long-term mortality among both men and women. In adjusted models, participation in physical activities had no significant association with mortality risk for men. Among women, participation in outdoor activities and leisure activities remained significantly associated with longterm survival in adjusted models. Among men, the benefits of physical activities for mortality risk were explained through demographic, health and psychosocial variables. However, participation in outdoor and leisure physical activities was beneficial for women and had a significant association with survival, even when controlling for demographic, health and psychosocial variables. Therefore, encouraging participation in specific physical activities, such as gardening, swimming and dancing among older women may improve long-term survival. 14 Item Health Index (1985) 555.95 (539) 461.84 (325) 404.02 (143) SAD Score 531.41 (515) 463.99 (320) 399.69 (144) Social Engagement Score 422.40 (516) 530.89 (319) 639.28 (143) Contributions WA, OF and PB all contributed to the design of this specific study, data analysis and the writing of the paper.