The purpose of this study was to identify dietary factors related to the incidence of diabetes mellitus in Korea. The study consisted of 165 diabetic patients, male and female, aged 30 to 70 years and 198 healthy persons as controls. Diabetic patients who had been diagnosed with diabetes mellitus for less than five years before the study period were recruited from eight different hospitals located in Seoul, Korea. Socioeconomic status, state of illness, physical activity, food habits and food intake were assessed. Food intake was assessed by food frequency questionnaire method using a 105-food frequency questionnaire developed for diabetic patients. The stress and activity indices of diabetic patients were not significantly different from control, but alcohol consumption and smoking levels were higher in diabetics than controls. Other diseases of male diabetic patients included liver diseases, digestive system diseases, and hypertension, while those of female diabetics were hypertension, neuralgia, arthritis and digestive system diseases. These disease patterns are different from Western countries whose most common complications are hypertension and hyperlipidemia. More irregular and less varied meals were found in the diabetic group compared to the control group, suggesting that diabetic patients have generally undesirable food habits. Otherwise, food and nutrient intake of diabetics did not differ greatly from the control group. It was found that diabetic patients consumed more cereals and less fruit than the control group, and also that male diabetics consumed more alcohol. The carbohydrate : protein : fat energy distribution ratio was 61.7 : 15.8 : 22.5 in male patients and 65.1 : 14.9 : 19.7 in female patients. Discriminant analysis showed that diabetes risk factors differed with sex. In male patients, the important factors were body mass index(BMI), vitamin C intake, family diabetic history and vegetable intake, while in female patients they were BMI, cereals intake, carbohydrate intake, vitamin C intake, stress, food habits and Ca intake. These results show that excessive intake of energy and fat are not the major causal factors in Korean NIDDM. Therefore, the diabetes risk factors of Western countries may not directly apply to Koreans. Mors study is needed to clarify the risk factors of Korean NIDDM.