Aiming at the problem that the slope is easy to lose stability in the mining of ionic rare earth mines, and the traditional stability criterion method is difficult to quantify the uncertainty of dynamic multi-field coupling, a slope stability evaluation model based on game theory and cloud model was constructed. According to the four principles of uniqueness, purpose, feasibility and observability, a multi-dimensional evaluation index system covering geomechanical parameters, leaching process, slope geological conditions, hydrometeorological characteristics and regional geological characteristics was constructed. The game theory was used to optimize the subjective weight determined by the fuzzy analytic hierarchy process (FAHP) and the objective weight determined by the entropy weight method, so as to realize the collaborative weighting of subjective and objective weights. The cloud model was introduced to quantify the membership degree of slope stability, and the expectation, entropy and super entropy were used to characterize the random fuzziness of parameter values. An ionic rare earth mining area in southern Jiangxi was taken as an example for industrial application. The results show that the comprehensive stability scores of 1^#, 2^# and 3^# slopes in the mining area are 74.8 (sub-unstable), 58.9 (relatively stable) and 99.3 (unstable), respectively, which are consistent with the actual observation results and the calculation results of the improved AHP method. Compared with the traditional limit equilibrium method, the potential risks can be found in advance. It is verified that the game theory-cloud model can provide reliable discriminant results and has high-risk sensitivity. The research results provide a theoretical basis for the dynamic control of slope stability of ionic rare earth mines.