Objective: This study investigated the level and correlates of self-care practices among patients with type 2 diabetes on follow-up in two public hospitals in Harar, Eastern Ethiopia. Methods: We conducted a hospital-based cross-sectional study on adult patients with type 2 diabetes, surveying diabetes self-care practices using a 15-item Summary of Diabetes Self-Care Activities. Responses ranged from 0 to 7 days, and a composite score was computed representing the mean days of diabetes self-care practices. A generalized Poisson regression model with robust variance was used. The association between the diabetes self-care practices and correlates was examined using the incidence rate ratio with a 95% confidence level. The statistical significance was set at a p value of ⩽0.05. Results: This study included 879 patients with type 2 diabetes. The overall mean (standard deviation) diabetes self-care practices were 3.7 ± 1.1 days out of the recommended 7 days, indicating low self-care practices. After controlling for other variables, tertiary educational level (incidence rate ratio = 1.06; 95% confidence interval: 1.01, 1.12), adequate diabetes knowledge (incidence rate ratio = 1.04; 95% confidence interval: 1.00, 1.08), moderate (incidence rate ratio = 1.07; 95% confidence interval: 1.02, 1.11) and high perceived self-efficacy (incidence rate ratio = 1.14; 95% confidence interval: 1.09, 1.13) (incidence rate ratio = 1.07; 95% confidence interval: 1.02, 1.11), high to marginal food security (incidence rate ratio = 1.13; 95% confidence interval: 1.03, 1.24), and receiving dietary advice (incidence rate ratio = 1.11; 95% confidence interval: 0.06, 1.15) were positively correlated with diabetes self-care practices. A history of hospitalization, on the other hand, was found to be inversely correlated with diabetes self-care practices (incidence rate ratio = 0.94; 95% confidence interval: 0.88, 0.99). Conclusion: The study indicated that adherence of patients with type 2 diabetes to the recommended self-care practices was considerably low. Therefore, tailored diabetes self-management education to enhance self-efficacy and diabetes self-care practices must be in place. This can be achieved through the system or individual-based integrated intervention efforts.