The number of diabetic ESRD patients has increased and death rates of diabetic patients on hemodialysis (HD), peritoneal dialysis (PD) and renal transplantation (RT) have remained higher than the death rate of non-diabetic patients. An attempt was made to compare the clinical characteristics, patients' cumulative survival, and technical survival among the three groups retrospectively according to the mode of renal replacement therapy(RRT), and to analyze the risk factors associated with mortality. A total of 229 diabetic ESRD patients diagnosed between 1986 and 1995 at the Severance Hospital who began dialysis or who underwent a kidney transplant were included and their medical charts were reviewed. Hypertension was the most common co-morbid disease in all study groups. The prevalence of cardiovascular disease was the only co-morbid condition that was significantly different among the three groups, which was highest in the PD group (24.4%) and lowest in the RT group (8%). In the analysis of a patient's cumulative survival rate not adjusted for age and sex, the RT group had the highest survival rate, and the cumulative survival rate of the HD and PD group were similar. The 5-year survival rate of the patients treated with HD, PD and RT was 28.8%, 19.8%, and 72.0%, respectively. No differences were observed in the patient's cumulative survival rate between the HD and PD patients even when it was adjusted for age. When adjusted for age, sex and risk factors, the relative death rate of the RT group was significantly lower in male patients younger than 60 years of age. With the exception of male patients younger than 60 years of age, the PD group showed a slightly lower relative death rate although it was not significant. The multiple Cox regression analysis of patient survival showed that age, serum albumin, BUN, mean hospital days, the presence of cardiovascular disease at the initiation of RRT were associated with mortality. The analysis of the technique survival rate revealed a better result in the HD group compared to PD group, but a limitation in being able to investigate the AVF function disturbed the accuracy of the analysis of technical survival rate. In conclusion, the survival rate between the PD and HD patients was not different and the RT group had the best survival rate. Therefore, kidney transplantation in diabetic ESRD patients should be considered positively if no other contraindicated condition for RT exit.