The invention discloses a brain-controlled robot system based on a multivariate synchronous exponential algorithm. The invention belongs to the field of brain control. The system comprises an electroencephalogram acquisition module, an online algorithm module, an interactive application module and a feedback module, wherein the online algorithm module comprises a signal calculation module, a calculation platform and a signal frequency identification module; wherein the interactive application module comprises a robot, a circulating receiving module and an interface building module for real-time and ordered instruction transmission. Sine coding stimulation is adopted in the design of a stimulation interface, and stable stimulation can be generated, so that a high-quality steady-state visual evoked potential is induced; an online algorithm adopts a multivariate synchronization index algorithm without training, and results can be classified in real time; an interface between the Matlab and the NAO robot is established by adopting a TCP/IP, and has the advantages of high real-time performance, order arrival of instructions and the like; and finally, all the adopted devices are wireless devices, so that the device is convenient to carry.