Many patients with end-stage renal disease are malnourished, and cross-sectional studies have shown that markers of malnutrition may predict death. In this study, we investigated the possible association of Subjective Global Assessment and mortality in a small cohort of Taiwanese hemodialysis patients. Fifty hemodialysis patients at a hemodialysis center in eastern Taiwan were enrolled in June 2002. Height and weight were used to determine the body mass index. Bioelectrical impedance analysis of body fat mass was performed before and after a mid-week dialysis session. Biochemical indexes of the nutritional status included serum albumin, creatinine, transferrin, cholesterol, and the normalized protein catabolic rate. Mortality data during 42 months after enrollment were obtained. Twenty-six hemodialysis patients were classified as well-nourished and twenty-four as malnourished based on Subjective Global Assessment. Decreased body mass index (p = 0.006), increased body fat mass (p = 0.019 before hemodialysis; p = 0.007 after hemodialysis), decreased serum albumin (p = 0.011), and decreased serum creatinine (p = 0.006) were significantly higher in the malnourished group. Older age (p = 0.042), decreased serum albumin (p = 0.028), decreased serum transferrin (p = 0.041), and malnourishment (p = 0.004) were significantly higher in the mortality group. Multivariate forward stepwise linear regression analysis of mortality and nutrition profiles show that Subjective Global Assessment is the independent predictor of mortality (R(2) = 0.20). Malnourished hemodialysis patients had a higher mortality rate than well-nourished hemodialysis patients in Taiwan. Subjective Global Assessment of the nutritional status appears to be a simple tool for assessing the nutritional status of hemodialysis patients in long-term care. This assessment tool is also beneficial for hemodialysis patients who are at a greater risk of nutritional-associated mortality.