Abstract: Based on the multi-factor influences of gas occurrence and field practice, combined with theoretical research on gas occurrence, a synergistic optimization method integrating Particle Swarm Optimization (PSO) and Simulated Annealing (SA) algorithms was employed to further enhance the accuracy of gas content prediction. This approach enables the PSO algorithm to have a higher probability of escaping local extremum points while improving the convergence speed of the SA algorithm. By combining the global search capability of the PSO algorithm with the local search capability of the SA algorithm, the limitations of traditional methods in gas content prediction are effectively overcome. A SAPSO-BP neural network model was established to predict gas content in unknown areas using existing gas content data. This model integrates the advantages of the PSO and SA algorithms, as well as the learning and fitting capabilities of the BP neural network, enabling more accurate predictions of coal seam gas content. The results indicate that the gas content in the sixth, eighth, and tenth panels of the No. 2 coal seam in Huangling No. 1 Coal Mine ranges from 1.54～5.40 m³/t, 1.86～3.70 m³/t, and 1.45～4.41 m³/t, respectively. The prediction accuracies of the BP, PSO-BP, SA-BP, and SAPSO-BP for gas content are 0.317, 0.593, 0.129, and 0.957, respectively, with the SAPSO-BP gas content prediction model being the most accurate. Using the SAPSO-BP gas content prediction model, the gas content in unknown areas of the mining area was predicted, resulting in a contour map of the mine's gas content.