In this study, the relationship between acoustic emissions and hardness of different rock types (model quartz, model calcite, and real iron ore) coupled with binary mix ratios of model quartz and iron ore (1:3, 1:1 and 3:1) was investigated in a laboratory-based AG/SAG mill. The acoustic emission response and sensitivity of the mill were compared, along with its product particle size distribution. Features, such as discrete wavelet transform (DWT), power spectral density estimate (PSDE), and statistical root mean square (RMS) were extracted from the mill acoustic emissions. From the results, it was evident that the mill acoustic emission response can be used to classify different rock hardness and their binary mixtures. Model quartz emitted the highest acoustic response, followed by iron ore and model calcite at the initial stages of grinding (5 min). The results further indicated that as the rock feed sizes reduced, the average mill noise also increased as a function of grinding time. Accordingly, the acoustic emission presented a contrasting effect with model calcite and model quartz emitting the highest and lowest noise emission, respectively. The proportions of different mineral types (model quartz and iron ore) were reflected well in the acoustic emissions. The study has demonstrated that the integration of acoustic sensing techniques in AG/SAG mills can serve as a proxy for online measurement of different rock/ore hardness, for example, as a fast track method to determine the hardness of ores in comparison to the traditional ore grindability procedures, such as Bond work index (BWI) and SAG power index(SPI) tests which can be laborious and time-consuming.