Health-related quality of life (HRQOL) measures predict cause-specific mortality, but few studies have explored whether generic self-reported HRQOL measures are independently associated with mortality in community-dwelling older persons. We postulated that a general measure of HRQOL, the short form 36-item questionnaire (SF-36), would be independently predictive of mortality among community-dwelling older persons. To evaluate this hypothesis, we followed a fixed cohort of 4,424 community-dwelling older persons recruited from a 2000 population-based survey in Taiwan until 2003 and investigated whether HRQOL was predictive of 3-year mortality, even after adjusting for traditional clinical risk variables. The data were collected via a door-to-door survey, and interviewers collected information on the subjects' demographics, medical history, utilization of health services, functional ability, falls, and self-reported physical and mental symptoms. Of the 6053 eligible subjects, 4,424 residents agreed to participate in the baseline survey and were contacted in 2003. During the 3-year period, the 3-year cumulative mortality rate for the study population was 5%. Mortality was significantly higher among males (5.57% vs. 4.27%, p = 0.049), and cumulative mortality increased with age (chi (2)-test for trend; chi (2) = 7.734, p = 0.001). For all scales except bodily pain, there was a significant relationship between a 10-point lower baseline score and mortality. Our primary multivariate risk model, which included two summary measures of HRQOL and significant clinical variables, demonstrated that a 10-point decrease in either the baseline Physical Component Summary (PCS) score or the baseline Mental Component Summary (MCS) score was associated with higher mortality (PCS: RR: 1.60, 95% CI: 1.39-1.83; p < 0.001; MCS: RR: 1.16, 95% CI: 1.01-1.34; p = 0.036). The findings suggested that low baseline PCS and MCS scores were important independent risk factors for 3-year mortality among community-dwelling older persons, even after adjusting for other risk factors.