<p>This study initially employs the discrete element software 3DEC to simulate the loading and unloading processes of instantaneous rockburst, encompassing stress application, boundary condition setup, parameter analysis, and model calibration. The accuracy of the simulation is validated through comparisons with experimental observations. Subsequently, leveraging the unique capabilities of 3DEC in localizing tensile and shear cracks, tailored algorithms for crack detection and optimization are developed to elucidate the dynamic evolution of internal and external fractures within rock masses at both macroscopic and microscopic scales. Furthermore, by integrating fracture dynamics and quantitative seismology theories, the multi-crack detection results are translated into the formation and evolution processes of seismic sources. Stress waves induced by multiple seismic sources at specific time points are calculated using theoretical formulas, followed by Fourier transformation and superposition, and compared with acoustic emission (AE) data to analyze their characteristic differences. A forward modeling approach (deriving stress or displacement fields from seismic sources) is employed to interpret the evolution mechanisms of internal cracks during rockburst. The complexity of rockburst mechanisms imbues AE curves with rich information content; however, current understanding of the variation patterns in these curves remains limited. This study aims to fundamentally explore the variation patterns of rockburst AE curves and to elucidate their complex failure mechanisms from the perspective of seismic source evolution.</p>