Thermal radiation is one of the main hazard characteristics of oil depot fire and the temperature field is the direct representation of the thermal radiation. Therefore, rapid and effective monitoring of the temperature field information is the best way to quickly acquire dynamic fire information. It is of significance to the improvement of emergent oil depot fire accident monitoring and fast response ability, the maintenance of people's life, property and safety, the protection of ecological environment and the great reduction of economic losses etc. On the basis of the research progress in ground surface temperature inversion methods at home and abroad, an oil depot fire temperature field information extraction model based on thermal infrared spectral and ground-based sensor data is established. By designing simulation experiments, the ground surface temperature conditions in the test side and its surrounding areas are inversed successfully. The result shows that this model can be used to quickly and effectively inverse the temperature field information of the area in which oil depot fire accident happened.