Satellite thermal infrared remote sensing has received worldwide attention in earthquake-precursors exploration. Meanwhile, it has also encountered great controversy due to the lack of quantitative interpretation of the observations, despite the existing qualitative physical mechanisms being able to greatly help us understand thermal infrared anomalies. Here, we report a potential mechanism to quantitatively analyze co-seismic thermal infrared anomalies based on temperature change caused by stress variation through theoretical, experimental, and field investigations. This paper firstly deduces theoretically the temperature variation during elastic deformation of rock on the basis of the thermodynamic theory. Secondly, three laboratory experiments on rock samples are conducted to verify the theoretical estimates of the temperature changes caused by stress variations using an infrared camera with the spectral range of 8~12 μm. Thirdly, a mechanical model on thrust faults is built to evaluate the co-seismic temperature drop as a result of thrust faulting. The model shows that the co-seismic temperature drop in rocks should be in the order of 0.18 K. This variation in rock temperature may cause a change in heat equivalent to changes in shallow atmospheric temperatures of 3.0–6.0 K, which is in accordance with the temperature anomalies observed by satellite thermal infrared remote sensing. In addition, the temperature change caused by crustal stress variation may involve a large spatial scale, covering the whole focal area, which has characteristics of regional distribution and is conducive to satellite observation. Therefore, a quantitative explanation of the satellite thermal infrared seismic anomaly mechanism can be given via the temperature change caused by crustal stress variation.