Abstract: With the gradual depletion of shallow coal resources in China and the increasing depth of mine excavations, the thermal hazard has intensified significantly. The tunnelling working site is a primary underground thermal hazard and requires targeted thermal hazard mitigation to ensure safe and efficient mine production. The premise of mine the thermal hazard control is to clarify its cooling load so that the great significance is to predict the air temperature in the mining operation space accurately. The airflow temperature prediction model of the tunnelling working site based on PSO-SVR was established, and the model was optimized by using the penalty factor C and kernel function parameter g in the model. Through field measurement and literature research, the training sample set of airflow temperature prediction in the tunnelling working site is established. By comparing with the MLR model estimated by the least square method and the conventional SVR model calibrated by the “trial and error” method, the advantages of the PSO-SVR algorithm are analyzed. The PSO-SVR algorithm model was applied to predict airflow temperature in J-24120 protective airway of Pingmei No.10 Coal Mine. Based on the prediction results of air temperature, the selection of refrigeration units and the design of cooling schemes are guided. The results show that: The PSO-SVR model has the best prediction performance, and the absolute error percentage of the model is only 1.85 %, which is 55.9 % lower than that of the conventional SVR model. So PSO optimization model parameters play an important role in improving SVR fitting degree, generalization and prediction accuracy. For every 100 m of roadway excavation, the average temperature rise of head-on airflow is 0.16 °C. When the roadway is excavated to 2 000 m, the temperature of head-on airflow in the roadway rises to 35.8 °C. Ji-24120 protective airway cooling demand is 1 083.28 kW, and the total cooling capacity of the design refrigeration unit is 1085 kW. After the cooling of the Ji-24120 protective airway, the average head-on temperature drop is 8.6 °C, and the cooling effect is remarkable, which shows the reliability and feasibility of the PSO-SVR prediction model of airflow temperature in the tunnelling working site.