Amputation is usually due to congenital defects, disease or injury. According to the Health Insurance Bureau of Taiwan, eight thousand amputation pieces have been done till 2003. Thus, a precise prosthetic is necessary. This research used the human body EEG to control bionic hand based on non-invasive electrodes for measuring brain motor area EEG, signal amplification and filter (hardware) and LabVIEW (software) for analyzing expectation movement state. Our results show that beta ERS characteristic wave could effectively interpret the action state and successfully self-control bionic hand in grasping movements.