Disclosed are a brain-computer interface based robotic arm self-assisting system and method. The system comprises a sensing layer, a decision-making layer and an execution layer. The sensing layer comprises an electroencephalogram acquisition and detection module and a visual identification and positioning module and is used for analyzing and identifying the intent of a user and identifying and locating positions of a corresponding cup and the user's mouth based on the user intent. The execution layer comprises a robotic arm control module that performs trajectory planning and control for a robotic arm based on an execution instruction received from a decision-making module. The decision-making layer comprises the decision-making module that is connected to the electroencephalogram acquisition and detection module, the visual identification and positioning module and the robotic arm control module to implement the acquisition and transmission of data of an electroencephalogram signal, a located position and a robotic arm status and the sending of the execution instruction for the robotic arm. The system combines the visual identification and positioning technology, a brain-computer interface and a robotic arm to facilitate paralyzed patients to drink water by themselves, improving the quality of life of the paralyzed patients.