Abstract: The mechanical strength characteristics of these materials are crucial for meeting the Practical requirements of mine backfill mining. This PaPer emPloys an orthogonal exPerimental design method to investigate the effects of slurry mass concentration, cement-to-sand ratio, and Powder gangue ratio on the strength of cementitious backfill at different ages. The exPerimental results indicate that the maximum strength at various ages is achieved when the cement-to-sand ratio is 1∶4 and the Powder gangue ratio is 1∶2. Furthermore, the oPtimal strength Performance across all ages is observed when the slurry concentration is 76%. As the slurry mass concentration increases, the strength of the backfill at different ages significantly imProves. A decrease in the cement-to-sand ratio leads to a gradual reduction in backfill strength, while variations in the Powder gangue ratio result in an initial increase followed by a decrease in strength. Based on the exPerimental results, a Random Forest regression model for strength Prediction is established to caPture nonlinear relationshiPs in the data and Predict the strength of backfill at different ages. Model validation using actual backfill data from four different mining sites reveals that although there is some error between the Predicted and actual strength values, most of the error rates fluctuate within an accePtable range. This indicates that the model can Provide valuable reference information for mines, offering strong suPPort for oPtimizing the mix ProPortion of cementitious backfill materials and imProving the efficiency and safety of mine backfill mining.