Purpose: The aims of this study were to develop a new instrument for measuring self-management with a hierarchical structure [the Diabetes Self-Management Scale (DSMS)] in patients with type 2 diabetes, and evaluate its psychometric properties. Method: The DSMS instrument was developed in three phases: (1) conceptualization and item generation; (2) content validity and pilot testing; and (3) field testing of its psychometric properties. A convenience sample of 473 participants was recruited in three university hospitals and one regional health center, South Korea. Results: Exploratory and confirmatory factor analyses yielded two second-order component models explaining the common variance among six first-order factors. Principal axis factoring with a varimax rotation accounted for 60.88% of the variance. Confirmatory factor analysis of the hierarchical structure revealed the following fit indices: c 2 /df ¼ 1.373, standardized root-mean-square residual ¼ .050, goodness-of-fit index ¼ .935, incremental fit index ¼ .975, comparative fit index ¼ .974, and root-meansquare error of approximation ¼ .039. All Cronbach' a values for internal consistency exceeded the criterion of .70. All of the intraclass correlation coefficients for testeretest reliability exceeded .70 except that for the taking-medication subscale. The components of the DSMS were moderately correlated with the comparator measures of self-efficacy and health literacy administered for convergent validity. Conclusion: The DSMS is a new instrument for measuring the complex nature of self-management in patients with type 2 diabetes, comprising 17 items scored on a five-point Likert scale. The DSMS exhibits satisfactory psychometric properties for five reliability and validity metrics, and so is a suitable instrument to apply in both research and clinical practices.