This dataset provides empirically validated Kuznetsov (Kuz) and Rampling (Ram) model calibration parameters—specifically K, n, and A values—for 12 geologically distinct rock types, including granite, basalt, limestone, sandstone, quartzite, schist, gneiss, dolomite, rhyolite, andesite, coal measure rocks, and weathered tuff. Derived from field blast monitoring data (fragmentation analysis via digital image processing, burden/spacing measurements, and explosive energy input tracking), it bridges the gap between theoretical Kuz-Ram predictions and real-world fragmentation outcomes. The dataset significantly improves predictive accuracy for blast design, reducing reliance on generic default parameters and enabling site-specific optimization of drilling and blasting efficiency.