The invention discloses a multi-information coupling prediction method for the risk of geostress type rock burst. The method comprises the steps of: (1) testing a stress value of a coal body in a burst danger area, and obtaining a distribution status of stress gradient; (2) blasting or striking a coal wall on one side of the coal body in the burst danger area, receiving waves through a wave pickup device on the other side, and then carrying out CT (computed tomography) imaging to obtain a distribution status of wave velocity gradient; (3) due to the coupling property between the stress gradient and the wave velocity gradient, determining a stress gradient critical value of rock burst occurring to the coal body, and obtaining a wave velocity gradient critical value VGm; (4) in a mining process of a coal mining machine, obtaining the wave velocity gradient VG of the coal body in real time by using a formula; and (5) comparing the wave velocity gradient VG obtained by real-time detection with the wave velocity gradient critical value VGm; if VG is smaller than VGm, determining that a prediction result is safe; and if VG is greater than VGm, determining that the prediction result is dangerous. The risk of the rock burst can be predicted in real time, thereby fully obtaining outburst risk information of the rock burst.