Abstract: The probability integral model is an important mathematical model to analyze the law of surface deformation caused by coal mining. The key parameters of mining subsidence can be predicted through the probability integral model. How to quickly and accurately obtain the prediction parameters of mining subsidence has always been the focus of scholars. To overcome the limitations of traditional methods in this field, this paper proposes a probability integral parameter inversion model based on the improved hunter-prey optimization （IHPO） algorithm. Based on the standard hunter-prey optimization （HPO） algorithm, IHPO introduces some improved strategies such as cubic mapping initialization, lens imaging refraction reverse learning, and a forced switching strategy, which significantly enhances the swarm intelligence optimization ability of the algorithm. IHPO is applied to the inversion of probability integral parameters, and the inversion model of probability integral parameters based on IHPO is constructed. The simulation results indicate that the relative error of the inversion probability integral prediction parameters for IHPO is maintained below 1.54%, and the fitting error for these parameters does not exceed 3.32. This suggests that IHPO offers a more precise inversion outcome compared to HPO. In addition, the parameter inversion model of IHPO has good robustness, can resist the influence of gross error interference, random error interference and missing observation points, and has strong global search performance. In practical applications, at the 1414（1） working face of Guqiao Coal Mine, IHPO is used for parameter inversion. The maximum fitting error for the inversion results is no more than 8.92, with the fitting errors for parameters q, tanβ, b, and θ all below 0.50, demonstrating high accuracy. The average fitting error for subsidence and horizontal movement values predicted by IHPO is 93.99 mm, which fully satisfies the accuracy needs of the actual working face, confirming the model’s effectiveness and practicality in predicting coal mining-induced subsidence.