• An intelligent approach using monitoring while cutting (MWC) data enables real-time prediction of rock strength and cuttability, facilitating adaptive excavation.
• A database of 132 conical pick-cutting experiments was established, integrating cutting parameters, responses, and rock mechanical properties for model training.
• The genetic algorithm-optimized backpropagation neural network (GA-BP) achieved the highest accuracy for predicting uniaxial compressive and tensile strengths.
• The radial basis neural network (RBF) proved most effective for classifying rock cuttability, combining AHP and fuzzy comprehensive evaluation.