Abstract: The initiation delay time is one of the key parameters in engineering blasting and plays a significant role in regulating the fragmentation size. Accurately predicting the fragmentation size under different initiation delay times is of great importance for guiding blasting engineering. To investigate the influence of initiation delay time on fragmentation size and to achieve fragmentation prediction, field tests with different inter-hole delay times were conducted in the graded aggregate blasting operation at the Lianghekou Hydropower Station. The fragmentation size distribution characteristics under different delay times were obtained. The results indicate that the effect of initiation delay time on fragmentation size exhibits zonal characteristics: the variation in small-sized fragments is not obvious, while the mean fragment size changes significantly, initially decreasing and then increasing with the increase in initiation delay time. The optimal delay per meter is approximately 3.3 ms/m. On this basis, numerical simulations of different inter-hole delay times were carried out using the finite element method, and the results were compared with the experimental data. The variation patterns obtained from both methods were in strong agreement. Analysis of the stress wave propagation characteristics under different initiation delays reveals that the intrinsic mechanical mechanism affecting the mean fragment size is the superposition effect of stress waves caused by different inter-hole delays and the reflection from new free surfaces. Finally, based on the fragmentation size distribution characteristics under different delay times, mathematical relationships between delay time and both mean fragment size and uniformity coefficient were established. A time term was introduced into the KUZ-RAM model, and a blast fragmentation prediction model considering delay time was proposed. The modified KUZ-RAM model overcomes the limitation of existing fragmentation prediction models in accounting for the influence of delay time on fragmentation size, thereby improving the accuracy of fragmentation size distribution prediction under electronic detonator delay initiation conditions.