•Opioid prescriptions were higher for those with diabetes vs without diabetes.•Opioid prescriptions decreased between 2008 and 2019, regardless of diabetes status.•Complications accounted for higher opioid prescriptions in those with diabetes. To determine the prevalence of opioid prescriptions among U.S. Medicare beneficiaries by diabetes status, and predictors of opioid prescription among those with diabetes. This retrospective study used claims data from the Centers for Medicare and Medicaid Services among beneficiaries age ≥ 65 years who were continuously enrolled in Part A, Part B, and Part D Medicare between 2017 and 2019 (N = 709,374). Logistic regression was used to determine the odds of opioid prescription among those with vs without diabetes; and, among those with diabetes, significant predictors of opioid prescription. Overall, the prevalence of any opioid prescription was 30.8 % among persons with diabetes and 24.2 % in those without diabetes (p < 0.001); chronic use was 8.0 % and 7.4 %, respectively (p < 0.001). Those with diabetes had a 45 % higher odds of having an opioid prescription compared to those without diabetes after adjusting for sociodemographic characteristics (OR = 1.45, 1.44–1.47). After adjustment for comorbidities/complications, the association reversed (OR = 0.83, 0.82–0.84). Persons with diabetes who had hypertension, obesity, CVD, neuropathy, amputation, liver disease, COPD, cancer, osteoporosis, depression, or alcohol/drug abuse had a 20 %-140 % higher odds of opioid prescription compared to those without these conditions. Comorbidities and complications accounted for the higher odds of opioid prescriptions among those with diabetes.