 This paper proposes a virtual reality AVR, based motor imagery, MI, a training system for post-stroke rehabilitation. It combines EEG signal mapping and EMG feedback to measure brain activity and provide real-time feedback of performance scores. The results show that the combination of MI and action observation AO produces the most pronounced event-related desynchronization ERD in alpha band and the highest muscle strength in beta band. This suggests that the VR-based MI training system has the potential to improve post-stroke rehabilitation. This article was authored by Mai Lin, Jin Li Huang, Jin Ming Fu, and others.