We sought to determine whether the MNA (Mininutritional Assessment) would be an effective tool for assessing the nutritional status of patients undergoing hemodialysis. Purposive sampling. Two hospital-managed hemodialysis centers in central Taiwan. Subjects were 95 ambulatory patients older than 40 years without acute diseases or infections who had received hemodialysis treatment at the center for longer than 30 days. Each subject was interviewed with a structured questionnaire to elicit basic personal data and health- and lifestyle-related information and answered questions on the Council on Nutrition Appetite Questionnaire and MNA. Serum biochemical data were obtained from their routine measurement. The nutritional status of each subject was graded with two modified MNA versions. MNA-TI adopted population-specific anthropometric cut-points, and MNA-TII further had the body mass index question deleted from the scale and question scores adjusted but maintained the same 30 total points. Based on the strength of correlation with the key nutrition-related parameters including appetite status, serum creatinine, percent weight loss, hospital length of stay, number of prescribed drugs, and hemodialysis time, both modified MNA versions reflected the nutritional status of Taiwanese hemodialysis patients better than the Council on Nutrition Appetite Questionnaire. Serum albumin did not reflect nutritional status well in these hemodialysis patients. MNA-TI predicted 26.4% and MNA-TII predicted 29.5% of hemodialysis patients were either malnourished or at risk of malnutrition. Both versions of the modified MNA are effective in assessing the nutritional status of hemodialysis patients. MNA-TII without body mass index is a significant improvement for hemodialysis patients because body weights fluctuate significantly between dialysis sessions.