Background The optimal choice of dialysis modality for diabetic patients remains controversial. This study aimed to compare mortality between peritoneal dialysis (PD) and hemodialysis (HD) in end-stage renal disease (ESRD) patients with type 2 diabetes (T2D). Methods Our observational, longitudinal cohort consisted of all incident ESRD patients with T2D who received either PD or HD in our center from January 2012 to December 2017 and were followed until December 2019. Propensity scores were used to select a 1:1 matched cohort. Mortality was compared between dialysis modalities using Kaplan-Meier survival analysis, and risk factors for mortality were estimated using multivariate Cox regression analyses. Results The median follow-up times were 35.5 months in the PD group (n = 134) and 41.6 months in the HD group (n = 134, p = 0.0381). The 1-, 2-, 3-, 5-, and 7-year patient survival rates were 98%, 91%, 77%, 61%, and 35% for diabetic PD patients and 96%, 88%, 81%, 60%, and 57% for diabetic HD patients. Kaplan-Meier survival analysis showed that overall mortality did not significantly differ between modalities (log-rank = 0.9473, p = 0.6575). Using a multivariate Cox regression model, advanced age and increased cholesterol at the initiation of PD treatment were independent risk factors associated with mortality, whereas under HD therapy, the risk factors associated with mortality were lower BMI and higher HbA1c. Conclusions These results suggest that in patients with T2D, mortality is comparable between PD and HD irrespective of whether there are the first 2 years or over the 2-year period, and that different mortality predictor patterns exist between patients treated with PD versus HD.