Exploring brain function is crucial for unraveling the pathological mechanism underlying stroke. While most studies focus on brain function and emphasize dynamic connections and interactions within or between brain regions, they often ignore the global properties of large-scale network topology. In this study, we analyzed resting-state electroencephalography (EEG) microstates in stroke patients, calculating key parameters such as mean duration (MD), occurrence (OC), time coverage (TC), and transition probability (TP) across different microstate classes. As a result, we identified four microstate classes (A-D) in both healthy subjects and stroke patients. Notably, stroke patients showed significant changes in Classes B-D, with increased MD, OC, and TC in Classes B and C and decreased MD, OC, and TC in Class D. In addition, stroke patients displayed higher TP between Classes B and C than healthy controls. Moreover, we observed a positive correlation between the OC of Class D and clinical Fugl-Meyer scores, suggesting a link between microstate dynamic and motor recovery. This study highlights the connection between specific microstates and brain regions, providing evidence that stroke patients demonstrate increased activity in visual and salience networks (SNs), but decreased activity in the dorsal attention network (DAN). We speculate these results from changes in the occipital lobe, frontal lobe, parietal lobe, and dorsal anterior cingulate cortex. These findings deepen our understanding of stroke pathophysiology and support further research.