More than 1,100 diabetes apps are available, but are infrequently used. A systematic review identified unsatisfactory diabetes app usability and its clinical effect to lower hemoglobin A1c level (0.15% to 1.9%), with variations in interactive app features for real-time feedback through automatic data analysis, clinician text messages, reminder alerts, or an app-initiated phone call. This result identified the need for health behavior theory applications to guide diabetes app usability evaluation. This study applied the Self-Determination Theory on human motivation to select app testing functions and to understand adult patient perspectives to use apps. A total of 92 adults with diabetes type 1 or 2 participated in a randomized crossover trial to test the usability of two top-rated Android diabetes apps (mySugr and OnTrack). Multivariable linear regression models assessed the effects of patient characteristics (i.e., age, education, and diabetes) and psychological needs on user satisfaction and user performance. Psychological needs important for motivation and behavioral change were associated with diabetes app usability. Higher user satisfaction was observed for participants who reported competence, autonomy, or connectivity with a healthcare provider. To enhance motivation to use apps for self-management, clinicians should consider addressing the patient’s competence, autonomy, and connectivity. User performance was associated with patient characteristics of age, sex, education, and diabetes duration because they affect the patient’s ability to use apps efficiently, successfully, and accurately. App training and ongoing technical support should be tailored for older adults, men, patients with less education, and those with diabetes duration more than 10 years.