The high geological stress field and high temperature observed at large depth result in a complex and uncertain engineering environment for developing deep mining projects. It severely restricts the safe and efficient mining of minerals and geothermal resources. A better assessment of the temperature and stress field is thus essential. Firstly, considering the uncertainty of subsurface parameters, the distribution of the temperature field will be evaluated based on numerical simulations and in-situ tests, using a new framework called Bayesian Evidential Learning. The most sensitive parameters for deep mining will be identified and the heat storage capacity of the ground will be evaluated. Secondly, a model of the wave velocity field in the mining area will be established by combining external active source and internal micro-fracture source of the rock mass. Based on this, the dynamic change the wave velocity field under the coupled action of geothermal and mining exploitation will be systematically analyzed. Finally, the evolution of the stress field deduced from temperature and wave velocity fields will be predicted, in order to forecast the risk of rock failure in the deep mining environment. This project will provide theoretical and technical support for the safe and efficient mining of deep minerals and geothermal resources.