The invention relates to an ore rock intensity prediction method based on component thermodynamic gene expression programming. With a provided component thermodynamic gene expression programming algorithm, the invention adopts a water absorption rate, a dry density, wave impedance, a dynamic Poisson's ratio, and a dynamic elastic modulus as input variables, adopts compressive strength as an output variable, automatically and intelligently digs out a mathematical model of the ore rock intensity, and thus predicts the compressive strength of the ore rock. The invention can quantitatively harmonize the balance between a selected pressure and population diversity in the gene expression programming, and thus improves the rate of convergence, solving precision and algorithmic stability of traditional gene expression programming when applied to the prediction of ore rock intensity.