<p>Predicting rock-cutting ability is one of the most fundamental issues in successfully implementing mining projects. The rock brittleness index (RBI) is closely related to rock-cutting ability and plays an important role in the cutting performance of drilling tools. This research is aimed at estimating RBI using the compressive strength ratio to tensile strength based on sedimentary rocks’ physical and dynamic characteristics. The data used were collected from mines in South China on limestone, dolomite, and sandstone. Feedforward artificial neural network (FFANN) based on different training algorithms, K-nearest neighbor regression (KNNR), classification and regression tree (CART), and multivariate linear regression (MLR) were used to estimate RBI. The influence of the texture of carbonate samples and lithology on geo-mechanical properties was investigated. The effect of sample texture on brittleness showed that limestone samples with mudstone texture have the least impact on brittleness. Compressional wave velocity and Schmidt hardness showed the greatest effect on RBI. The modeling results showed that the FFANN method, based on the spider diagram, the model’s performance index (MPI), and the non-parametric test, is more accurate than other methods with a determination coefficient of more than 99% and an MPI value of 1.91 to estimate the RBI. The results of this study will be used in future studies to predict drilling rates and rock burst phenomena.</p>