Decalcified bone matrix (DBM) is a widely used alternative material for bone transplantation. In the DBM production process, an effective particle size and the highest utilization rate of raw materials can be achieved only through multiple high-speed circulating comminution. The rat posterolateral lumbar fusion model (PLF) is the most mature small animal model for the initial evaluation of the efficacy of graft materials for bone regeneration and spinal fusion. To evaluate the differences in the in vivo osteogenic effects of DBM pulverization through 1, 5, 9, and 14 high-speed cycles, sixty athymic rats were divided into six groups: single cycling crushing (CC1), 5 cycles of crushing (CC5), 9 cycles of crushing (CC9), 13 cycles of crushing (CC13), autogenous bone graft (ABG) and negative control (NC). Posterolateral lumbar fusion was performed. Six weeks after surgery, the bilateral lumbar fusion of athymic rats was evaluated through manual palpation, X-ray, micro-CT and histological sections. Rank data were tested by the rank-sum test, and nonparametric data were tested by the Kruskal‒Wallis H test. The manual palpation and X-ray results showed that the fusion rate did not significantly differ between the CC1, CC5, CC9, CC13 and ABG groups. However, cavities appeared in CC9 and CC13 on the micro-CT image. The bone mass (BV/TV) of CC1, CC5, CC9 and CC13 was better than that of the ABG group, while almost no osteogenesis was observed in the NC group. Histologically, there was no obvious difference between the four groups except that the CC9 group and CC13 group had more fibrous tissues in the new bone. In conclusion, DMB with different cycling crushing times has no obvious difference in fusion rate of PLF, but it is slightly better than the ABG group.