The process of high-stage filling stability risk assessment is characterized with fuzziness and randomness. A high-stage filling stability evaluation model with the introduction of the cloud model theory is proposed in this paper. Taking four domestic deep mines including Anqing Copper Mine as the research objects, 16 influential factors, such as slump, lateral exposure area, filling slurry slope, and filling effect, were selected as risk assessment indicators. The cloud model parameters of each index belonging to different risk levels were calculated with the help of cloud theory. At the same time, based on the improved CRITIC-G1 method, the subjective and objective weights of the evaluation indexes were optimized and integrated into the comprehensive weights of the indexes. Then, the comprehensive certainty degree of the backfill stability was calculated by using the forward Gaussian cloud generator, and the stability grade of the filling body of four mines was finally determined. The results show that the risk levels of Anqing Copper Mine, Lilou Iron Mine, Dongguashan Copper Mine and Sijiaying Iron Mine are grade IV, grade IV, grade III, and grade II, respectively, and the filling body of Sijiaying Iron Mine is relatively stable. Compared with other mathematical models, the proposed model not only ensures the accuracy of the results, but also uses cloud numbers to represent the fuzziness and randomness so as to obtain the stability grade of the filling body under the coupling effect of different evaluation indexes. The model realizes the visualization of the classification, and has certain reliability, providing a new method for the risk assessment of the stability of high-stage filling body.