Abstract: Under the background of the rapid development of intelligent coal burst monitoring technology, how to use massive seismic monitoring information to investigate source rupture and apply it to the prediction, early warning, or prevention of coal burst is the focus and difficulty of future research. To solve the problems of insufficient investigation of seismic clustering fracture mechanism aggregation law and overcome the difficulty of direct application of seismic source mechanism inversion technology to the prediction and early warning of coal burst, the theoretical analysis and numerical simulation methods are used to establish a prediction method of coal burst risk based on seismic source mechanism and location error calibration. Taking the 2215 working face with frequent high magnitude seismic events in a coal burst mine in Inner Mongolia as an example, this paper summarizes the characteristics of the limited number of seismic waves, large noise impact, and high inversion efficiency. Also, it compares and analyzes the characteristics, advantages, and disadvantages of different moment tensor inversion methods, and establishes that the hybrid moment tensor inversion method is more suitable for coal mine application scenarios. The statistical results of the rupture mechanism of high magnitude events in the 2215 working face show that the main inducing events of roof rupture are a tensile shear failure and compressive shear failure. The shear component statistics of seismic events before the occurrence of high magnitude events show that the proportion is not more than 40%. The occurrence statistics of source rupture surface show that the strikes of seismic events dominated by roof breaking are mainly distributed in the range of 3°−80° and 150°−270°, and the dips are mainly distributed in the range of 70°−90°, with prominent agglomeration characteristics. Based on the common source rupture mechanism and location errors in the coal mine, the possibility of seismic event aggregation is analyzed, and the criteria of seismic aggregation are established. The risk prediction method and index based on the source rupture mechanism and location error calibration are constructed. The results of prediction examples show that the identification of seismic anomaly aggregation areas can be significantly enhanced by considering the source rupture mechanism and location error. When the seismic data is relatively complete, the I_P threshold of the risk prediction index based on the source rupture mechanism and location error calibration is set to 0.7, and the accurate prediction of high magnitude events of different rupture types can be achieved.