Land surface temperature (LST) is among the most important indices in the studies related to earth surface such as conservation, energy exchange and the water between the land surface and atmosphere. The main goal of this study is to present an algorithm in order to estimate the land surface temperature using the data of the Hyperspectral Thermal Emission Spectrometer (HYTES). The HYTES sensor has 256 bands in the range of 7.4-12 micrometers that bands in the range of 7.4-8 micrometers are removed due to strong water vapor in this spectral region and bands above 11.5 micrometers are removed due to issue of calibration. 202 bands remain, which we want in this study to obtain optimal bands from 202 bands using the genetic algorithm and then obtain the land surface temperature using those bands. We need to define a cost function and appropriate initial parameters for the genetic algorithm to select optimal bands. In this research, the cost function is to minimize the temperature difference between the thermal product of the sensor and the obtained land surface temperature with a split window algorithm and the number of variables in each gene is as large as the number of bands (202) and the initial population is 80 in genetic algorithm. The bands used in split window algorithm are selected using the genetic algorithm. In this study, we use split window algorithm that obtain land surface temperature through optimal bands that are selected using Genetic algorithm among 202 bands. Generally, in this study, first we use Genetic algorithm to choose optimal bands from 202 bands and obtain the coefficients of split window algorithm.The number of these coefficients is dependent on the number of bands which are selected by Genetic algorithm. Then, by using resulted coefficients that are obtained with least squares method and selected bands, land surface temperature is obtained for two different data through split window algorithm. In this research, a small part of the first data was used as training data for the genetic algorithm to obtain the coefficients algorithm of the split window and optimal bands to calculate the land surface temperature for the rest of the data. In a separate window algorithm, in addition to the algorithm coefficients, we need the emissivity of the relative bands used in split window algorithm. In this study, we used the emissivity product of the HYTES sensor. Among 202 bands, 110 bands are selected using the genetic algorithm. Using this 110 bands, split window algorithm coefficients and bands emissivity, land surface temperature is calculated for two data and evaluated. Finally, the thermal product of HYTES is used to evaluate the our proposed method and indicate its accuracy. The temperature obtained using proposed algorithm for both data is evaluated with reference data (thermal product) and the RMSE value is resulted as 0.025 and 0.999 for the first and second data respectively. Therefore, according to the obtained errors, we can argue that the proposed algorithm is an appropriate method to obtain the land surface temperature using the data of HYTES.