Abstract: To improve the prediction accuracy of the spontaneous combustion temperature of coal in goafs and accurately identify the combustion risks during the heating stage of the coal body, a prediction model (PLO-GBDT) for the spontaneous combustion temperature of coal in mined-out areas based on multi-parameter fusion is proposed. Programmed temperature experiments were conducted to collect data on the volume fractions of key indicator gases O₂, CO, CO₂, CH₄, C₂H₆ and C₂H₄ as temperature increased, and a multi-parameter dataset for coal spontaneous combustion in goafs was constructed. Through the analysis of feature importance, the indicator gases that significantly affect the ignition temperature of coal spontaneous combustion were identified. O₂, CO, CO₂, CH₄, C₂H₆ and C₂H₄ were selected as the input variables of the model. The polar light optimization (PLO) algorithm was used to optimize the parameters of the gradient boosting decision tree (GBDT) model. A PLO-GBDT model for predicting the ignition temperature of coal spontaneous combustion was constructed. The model used 70% of the sample data as the training set and 30% as the test set. The prediction performance of the model was evaluated using evaluation indicators such as MSE, RMSE, MAE, MAPE and R². The results show that the MSE, RMSE, MAE, MAPE and R² of the PLO-GBDT model prediction are 0.000 03, 0.005 50, 0.003 70, 0.010 70 and 0.979 40 respectively. The established model has high fitting accuracy and stability. To verify the superiority of the model, it is compared with five other models. The results show that the R² values of the PLO-GBDT model compared with the reference models PLO-MLP, PLO-CNN, PLO-GRU, PLO-Informer and PLO-XGBoost are 0.967 7, 0.659 7, 0.785 3, 0.846 1, 0.895 6 and 0.948 5 respectively.