BACKGROUND AND AIMS: The psychometric properties of the EQ-5D-3L (a generic preference-based instrument used for calculating quality-adjusted Life -years) have not been known for any type of disease in Iran. This study aimed to assess its validity and reliability in patients with type 2 diabetes. METHODS: Data of 579 patients were collected from the Diabetes Research Center and Clinics in Yazd using EQ-5D-3L, SF-36, and DQoL instruments. The ceiling effects were computed for the EQ-5D-3L index and EQ VAS. The construct validity was assessed by using convergent validity, discriminant validity, and known-groups validity. Reliability was assessed using Cohen's kappa value, Cronbach's alpha, and intra-class correlation coefficient. RESULTS: The ceiling effects of EQ-5D-3L and EQ VAS were 20.18% and 15.33%, respectively. The highest degree of correlation was found between the pain/discomfort of EQ-5D and the BP scale of the SF-36 (0.55). Higher scores of all scales of the DQoL were associated with patients reporting no problems in each EQ-5D dimension. The mean of EQ-5D-3L index and EQ VAS scores were significantly higher in the male, married, and employed patients, and they did not have retinopathy, nephropathy, IHD, hypertension, DFU. The range of kappa values was from 0.39 to 0.71, and value of ICC for the EQ-5D-3L index and EQ VAS was 0.76 and 0.64, respectively. Cronbach's alpha was 0.87 for EQ-5D-3L and 0.74 for EQ VAS. CONCLUSIONS: Our findings demonstrated good construct validity and moderate to good levels of reliability of the EQ-5D-3L for using in the patients with diabetes, and it can be used in research or clinical practice in Iran and other regions of the Middle East.