The invention discloses an OpenVIBE-based high-performance motor imagery online brain-computer interface system, including signal acquisition equipment, signal test scripts, signal acquisition scripts, data training scripts and online experiment scripts, wherein the signal test script is connected to the signal acquisition equipment, Detect the signal quality through the signal test script, then set the experimental parameters through the signal acquisition script and collect the data of the motor imagery experiment, then use the data training script to realize the training of the spatio-temporal filter classifier based on the RSTFC algorithm, and obtain the specific spatio-temporal filter classifier to import The online experiment script is described, and the online experiment script implements a high-performance online brain-computer interface system for motor imagery based on the trained spatio-temporal filter classifier. The invention adopts a modular design method to improve the readability and flexibility of the system, is convenient for function expansion, greatly improves the work efficiency of researchers, and has the advantages of high accuracy and good performance.