Objective: There has been increasing awareness that micro-vascular alteration or vascular inflammation has been associated with levodopa-induced dyskinesia in PD. Vascular endothelial function assessed by flow mediated dilation (FMD) is known to reflect early microvascular change. We compare the impact of levodopa or dopamine agonist treatment on the change of FMD in de novo PD patients. Methods: This retrospective study used a selected sample from registry. We identified de-novo PD patients who underwent FMD at baseline, and follow-up FMD after 1 year (± 2 month) of levodopa (n = 18) or dopamine agonist (n = 18) treatment. Results: FMD decreased after levodopa (8.60 ± 0.46 to 7.21 ± 0.4, p = 0.002) but there were no significant changes after DA treatment (8.33 ± 0.38 to 8.22 ± 0.33, p = 0.26). Homocysteine rose (11.52 ± 0.45 to 14.33 ± 0.68, p < 0.05) during levodopa treatment, but dopamine agonist had no effect (10.59 ± 0.38 to 11.38 ± 0.67, p = 0.184). Correlation analysis revealed that the changes in homocysteine level had non-significant correlation with FMD change (r = − 0.30, p = 0.06). FMD change was not associated with age (p = 0.47), disease duration (p = 0.81), baseline motor UPDRS (p = 0.43), motor UPDRS change (p = 0.64), levodopa equivalent dose change (p = 0.65). Conclusions: We found that 1-year levodopa treatment may adversely affect vascular endothelial function in de novo PD. Further studies are needed to clarify the exact pathogenesis and clinical implication of levodopa-induced endothelial dysfunction in PD.