Objective: To investigate the survival rate and to estimate their hazard ratio of death in the post-stroke elderly in the communities of Beijing. Methods: The study samples were collected from districts (or counties), subdistricts (or townships) and residents committees (or natural villages) in 1997 using stratified classification, multistage random sampling according to the principles of cluster sampling. The individuals were questionnaired. A total of 2788 valid samples were obtained. These samples were followed up comprehensively in 2000 and 2004, respectively; the time and reasons of death and lost to follow-up were documented. Results: The year mortality in stroke group was 7.77/100 person-years, and in non-stroke group was 4.51/100 person-years (χ2 = 27.69, P < 0.001). The difference was statistically significant. The cumulative survival rate was 57% in the stroke group up to the end of the follow-up period, and it was 72% (u = 29.24, P < 0.001) in the non-stroke group. The difference was statistically significant. With the extension of time, the cumulative survival rate in the stroke group was significantly lower than that in the non-stroke group; the survival rates of 1, 2, 3, and 5 years in the stroke group were 86%, 78%, 73%, and 64%, respectively. They were all lower than 94%, 89%, 85%, and 78% in the non-stroke group. After adjusting for age, sex, hypertension, coronary heart disease, diabetes, tumor and other chronic diseases with Cox regression, the mortality risk ratio of stroke was 1.869 (95% CI: 1.530-2.283). The attributable risk (AR) was 3.95, and the attributable risk proportion (ARP) was 46.5%. Conclusions: The survival rate of the post-stroke elderly in the communities was significantly decreased. Stroke is one of the important factors of the death in the elderly. So the whole society should pay attention to the prevention of stroke.