<p>Rock fragmentation prediction in mining is often limited by insufficient integration of geological structure into conventional blast design models. This study develops an integrated framework that combines quantitative crack indices with the Geological Strength Index (GSI) to improve prediction and control of blast-induced rock fragmentation in open-pit iron ore mines. Field data from 90 production blasts conducted across four mechanized mines in Banded Iron Formation (BIF) were analyzed to characterize fracture networks using crack density and crack saturation derived from core logs and bench observations. These parameters were systematically integrated with GSI-based rock mass assessment to establish a unified predictive and classification approach. A multivariate logarithmic model demonstrated substantially improved agreement with measured fragmentation compared with traditional empirical methods. Based on the integrated geological characterization, a five-zone classification system was developed to guide powder factor adjustments according to rock mass conditions. Field validation trials confirmed that zone-based optimization enhances fragmentation consistency in intact zones while enabling significant reductions in explosive consumption in highly fractured zones without compromising operational requirements. The proposed framework provides a practical pre-blast decision-making tool that directly links geological characterization with blast design parameters. This approach supports more efficient, economical, and geology-responsive blasting practices and offers a transferable methodology for mining operations in structurally complex rock mass.</p>