Diabetes increases the risk of Parkinson disease (PD). Sodium-glucose cotransporter 2 (SGLT2) inhibitors, a new glucose-lowering therapeutic class, have shown neuroprotective effects in mechanistic studies. However, the association between SGLT2 inhibitors and PD risk in real-world populations with type 2 diabetes (T2D) remains unclear. The aim was to assess the association between SGLT2 inhibitors and the risk of PD in older populations with T2D. This retrospective cohort analysis used Medicare claims data from 2016 to 2020 to identify fee-for-service beneficiaries ≥65 years diagnosed with T2D and without pre-existing PD. The initiation of an SGLT2 inhibitor was compared with that of a dipeptidyl peptidase-4 (DPP4) inhibitor. The outcome was the first incident PD ever since the date initiating either an SGLT2 inhibitor or a DPP4 inhibitor. We employed a 1:1 propensity score matching to balance the baseline covariates between treatment groups, including sociodemographics, comorbidities and co-medications. We applied Cox regression models to assess the effect of SGLT2 inhibitors versus DPP4 inhibitors on incident PD. Of 89 330 eligible Medicare beneficiaries (mean age: 75 ± 7 years, 52% women), 0.6% (n = 537) had incident PD over the follow-up. After 1:1 propensity matching, the PD incidence was 2.5 and 3.5 events per 1000 person-years in the SGLT2 inhibitor group and DPP4 inhibitor group, respectively. The SGLT2 inhibitor group was associated with a significantly lower risk of incident PD than the DPP4 inhibitor group (hazard ratio: 0.70 [95% confidence interval: 0.55-0.89]). There is a potential trend that the risk reduction in incident PD was profound in non-Hispanic Black individuals and insulin users. Compared to DPP4 inhibitors, SGLT2 inhibitors were associated with a significantly lower risk of incident PD in older populations with T2D.