The correct use of the product is possible only when the land surface temperature （LST） data is calculated by an accurate and reliable inversion algorithm. In this paper， we compare the inversion results of five commonly used LST inversion algorithms based on Landsat-8， Landsat-9 data， and weather station data. The inversion results and parameter sensitivity analysis of different algorithms are tested. The results show that the inversion results of the Radiative Transfer Equation （RTE） and Single Channel （SC） algorithms calculated based on land surface emissivity （LSE） are in good agreement with the ground measured. The inversion results of the SC algorithm based on the atmospheric water vapor inversion and the Split Window （SW） algorithm based on the atmospheric water vapor inversion are higher than the measured temperature. The inversion accuracy of the Mono Window （MW） algorithm based on average temperature parameters is not ideal. In addition， the consistency of the inversion temperature of the two data on different ground objects is compared. Our study can provide a reference for land surface temperature inversion based on Landsat-9 data.