Abstract: Ai ming at the co mplex fracture develop ment in overburden and repeated surface disturbance under multi-sea m repeated mining, focuses on shallow-buried multi-sea m working faces. Based on DEM si mulation, fractal geo metry is e mployed to quantitatively characterize overburden fractures, and fracture rate theory is introduced to clarify the evolution of fracture develop ment and surface da mage in coal sea m groups. The results show: ① As mining progresses downward, fractures undergo a cyclic evolution process of “activation–expansion–co mpaction–reactivation.” Under close-distance mining, the overburden breakage angle maintains a consistent trend, while the disturbance effect weakens with increasing sea m spacing, leading to divergent breakage angles and a “double-trapezoid” fracture morphology. ② Fractal geo metry theory reveals the multi-disturbance and zonal evolution characteristics of fractures in multi-sea m mining. Global fractal analysis indicates that after the first sea m extraction, the fractal di mension exhibits steady growth (average incre ment: 0.06) under various factors, whereas subsequent extractions cause an initial decrease (fracture closure) followed by expansion. Notably, larger sea m spacings exhibit stronger disturbance effects on fractal di mensions co mpared to close-distance mining. Zonal fractal di mensions further classify fracture develop ment into three regions: “fracture—developed zone,” “disturbance-activated zone,” and “co mpaction-closed zone.” ③ I mage processing techniques were e mployed to extract fracture lengths, showing an average length of 39.4 m for fractures at depths <100 m, significantly higher than those below 100 m (23.7 m). Fracture rate analysis de monstrates pronounced incre mental growth (21.26–27.42 m sea m spacing) but attenuated disturbance effects at 86.8 m spacing, coinciding with increased surface subsidence. Fractures dyna mically alternate between propagation and closure during mining. ④ Surface subsidence predictions, integrating si mulation results, maxi mu m subsidence models, and the probability integral method, achieved absolute errors ≤3.6%. A positive correlation was identified between surface subsidence and fracture develop ment in shallow multi-sea m mining.