[en] One of the key outcomes of blasting in mines is rock fragmentation that profoundly affects downstream expenses. In fact, size prediction of rock fragmentation is the first step towards the optimization of blasting design parameters. This paper attempts to present a model to predict rock fragmentation using Mutual Information in Meydook copper mine. Ten parameters are considered to influence fragmentation. On the other hand, Rock Engineering System is employed in order to compare different models. To validate the results, six blasting scenarios were selected and the results were compared. The coefficient of correlation (R2), Root Mean Square Error and Mean Absolute Error were used to assess the performance of presented models. The R2, Root Mean Square Error and Mean Absolute Error values for 30 blasting cycles were calculated to be 0.81, 10.7, and 9.02 for Mutual Information model, and 0.75, 11.87, and 9.61 for Rock Engineering System, implying the better capability of Mutual Information model to predict fragmentation.