The average dip angle and thickness of a silver ore body are 70 degrees and 10 m respectively.The ore body is unstable and the surrounding rock of the hanging wall and footwall wall is relatively stable.In order to determine the optimum stope structure parameters and optimize the safety and economy of the mining scheme，a Vague-RSM-AFSA model was established to optimize the stope structural parameters.Fifteen schemes of stope structural parameters were designed based on the principle of central composite test method，and the numerical model of each scheme was established by Midas coupled FLAC^3D. Stope structural parameters are related to the safety and economy of mine production.The maximum settlement displacement of stope roof，maximum horizontal displacement of pillar，mining-cutting ratio and the ore loss rate were selected as evaluation indexes based on the geological conditions，mining methods and other factors.The Vague theory can accurately express the fuzziness and uncertainty of information，the weight of each index was obtained by calculating Vague entropy of each index.Based on Vague set，the positive and negative ideal schemes within the optimum range of stope structural parameters were obtained，and the superiority of each scheme was calculated by Euclidean distance method.Response surface methodology （RSM） can accurately construct the complex non-linear response relationship between independent variables and dependent variables，response surface method （RSM） was used to establish the response surface model of stope structural parameters and superiority of various schemes in central composite test.The square R ² of the fitting correlation coefficient of the model is 0.9766.The numerical analysis of three groups of structural parameters in the optimum range but not in the central composite test was carried out.The maximum error between the results and the response surface model is 0.0213，which shows that the established response surface model has high accuracy.Artificial fish swarm algorithms （AFSA） has a strong ability to search global extremum，and can effectively solve complex non-linear multi-objective optimization problems，the artificial fish swarm algorithm （AFSA） was used to optimize the superiority response surface model.The optimum stope structure parameters were obtained as follows： 20 m for stope height，32.8 m for stope length，16.1 m for pillar width.And the superiority is 0.2631，which is higher than the maximum superiority of 0.2271 in the central composite test，and the error between the optimization results and the numerical simulation results is 0.0037. It shows that the Vague-RSM-AFSA model has good optimization ability and high accuracy，which provides a new method for the optimization of stope structural parameters.