 This paper proposed a novel algorithm for estimating land surface temperature, LST, from thermal infrared sensor, TRS, data aboard Landsat 8. The algorithm utilized a modified split-window covariance-variance ratio method to determine the coefficients of the algorithm, which were then used to calculate the emissivity of the atmosphere and the fraction of vegetation cover. These values were then combined with the emissivity of the land surface to calculate the LST. The simulation results demonstrated that the algorithm could accurately retrieve LST with an error of less than 1K. Additionally, the algorithm's performance was compared to other existing algorithms and found to be superior. Finally, the algorithm was tested in various regions and proved to be reliable. This article was authored by Chen Du, Hua Zhongren, Qi Mingqin, and others.