Land surface emissivity (LSE) is an important intrinsic property of materials and knowledge of the LSE is essential to derive the land surface temperature (LST) that can be obtained from the emitted radiance measured from space. LSE provides useful information for geological and environmental studies, mineral mapping and is one of the important input parameters for climate, hydrological, ecological and biological models. The emissivity of natural surfaces inherently may vary significantly due to differences in soil structure, soil composition, organic matter, moisture content and differences in vegetation cover characteristics. In other words, LSE changes is depending on the surface (such as texture, topography, soil moisture, angular variations effect) and sensor parameters (such as spatial resolution, SRF, and effective wavelength of thermal bands). Remote sensing technology provides widely the monitoring of this quantity. Several methods exist to estimate LSE from satellite data, which apply the visible and near-infrared (VNIR) or thermal infrared (TIR) spectral regions or both of them. According to the way by which the LSE is determined along with LST, the emissivity estimation methods from optical remote sensing data can be categorized into three distinct types including, stepwise retrieval methods, simultaneous LST and LSE retrieval methods with known atmospheric parameters, and simultaneous LSEs, LST, and atmospheric quantities retrieval methods. Influential researches, in the stepwise retrieval methods, were conducted and mainly NDVI-methods have been used to predict LSE from NDVI values. In particular, NDVI-methods assume that the surface is composed of the soil and vegetation, some problems arise for other kinds of surfaces that are likely classified as bare soil pixels, such as rocks, man-made, and ice/snow. Besides, another main origin of error in these methods is caused by great changes in the emissivity of soil types. Furthermore, the choice of a typical emissivity value for some surface objects such as bare soil is a more critical question, because the variability of emissivity values for soils is more than vegetation and other ones. In this research, a new approach called improved normalized difference vegetation index-based method (INDVI_based) estimating LSE on Landsat-8 (known as Landsat Data Continuity Mission, LDCM) data has been proposed for semi-arid areas. At first, a simulation of channel emissivities and reflective bands of basic classes in bare soil, vegetation and mixed areas is accomplished based on convolving ASTER spectral Library with LDCM spectral response functions. Then, for main three areas are defined to determine separate emissivity estimate model as function of reflective bands from basic spectra associated with the main class. In the proposed method, the cannel LSEs are expressed as functions of atmospherically corrected reflectance from the LDCM visible and near-infrared channels with wavelength ranging from 0.4 to 2.29 μm fo bare soil. The effectiveness of the proposed approach was implemented in LDCM data and obtained LSE were compared and validated with two scenes of LSE standard product of Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER). Results showed that LSE of the improved proposed method, in the band 10 of LDCM in comparison with the first and second LSE product of ASTER, lead to 0.76% and 0.75% errors in term of root mean square error (RMSE) measure, respectively. Moreover, this error for thermal band 11 is1.49 % and 1.06% in first and second examined scenes, respectively. Unlike previous methods, the proposed method not only accurately estimates of LSE  as a function from the reflectance of various surface objects, but also it  uses  the spectral response function  of thermal and reflective bands in estimating the LSE. In addition, the proposed method the poor relationship between LSE and only reflectance of the red band in previous methods, strengthen due to the use of all reflective bands in LSE estimation and it is applicable on most sensors.