Microseismic (MS) source location is a core technology in MS monitoring that helps to infer the mechanical state of the rock mass and prevent dynamic disasters. However, when considering P-wave arrival time system errors (PATSEs), conventional location methods encounter the problem of high inversion dimensions and do not consider the influence of large P-wave arrival time errors. Therefore, we proposed a low-dimension double event joint location (DEJL) method that considers the PATSEs and introduces data field theory to reduce the influence of large P-wave arrival time errors. First, the P-wave travel-time equation of double events was constructed to eliminate the earthquake occurrence time and PATSEs, and the high-dimension inversion was reduced to a set of equations containing only six parameters. Then, 2000 P-wave arrival time combinations from more than six sensors were selected, and a Newton iterative based algorithm was proposed to solve the location point of each combination. Finally, the location point with the largest potential value of the data field was considered as the location result. The Institute of Mine Seismology acquisition system, placed in the Yongshaba mine in Guizhou Province, China, was selected to conduct the tests. P-wave arrival time datasets were obtained from two synthetic locations and eight blasting events. The results showed that the average location error of the blasting events using the proposed method is 32.59 m, which is lower than that of the homogeneous velocity-based location results. Thus, we determined that the proposed method offers a good MS location accuracy and broad application potential.