In view of the continuous improvement of industrial automation, the online monitoring requirements for temperature are becoming much more widespread, and the traditional thermocouple temperature monitoring system based on look-up table and hardware compensation methods is difficult to adapt to the problem of higher accuracy requirements. In this paper, an online temperature monitoring method for thermocouples based on Gradient Boosted Decision Tree (GBDT) algorithm is proposed to improve cold junction compensation and nonlinear problems in online temperature monitoring of thermocouples. It can be seen from the experimental test results that the GBDT algorithm is applied to the online temperature monitoring of thermocouples, which can effectively improve the accuracy of online temperature monitoring and provide a new solution for industrial automation of temperature monitoring.