Abstract: In order to determine the relationship between thermodynamic entropy yield, cumulative entropy and fatigue performance of metal components during low cycle fatigue, the fatigue tester was used to perform low cycle fatigue cycle loading on Q235 specimens, and infrared thermal imager was used to record different loading amplitudes and the surface temperature of the specimen under the frequency is calculated by entropy yield and cumulative entropy, and then the relationship between fatigue life and entropy is constructed. The results show that under the same loading frequency, the entropy yield is proportional to the loading amplitude, the fatigue period is relatively low, the entropy value of fatigue accumulation is inversely proportional to the loading amplitude, and the accumulated entropy is only related to the loading amplitude. The loading frequency is irrelevant. Under the same loading amplitude, the entropy yield is proportional to the loading frequency, and the entropy curve accumulated in the fatigue process is linear with the loading time. At the same time, the fatigue life model was obtained by using the number of cyclic loading cycles and the cumulative entropy as parameters, and it was verified that the model could predict the fatigue life of metal components more accurately in the form of non-destructive testing.