Applying atmospheric corrections on satellite images is an important step before using the satellite image for any further processing. These atmospheric corrections are broadly classified as either image-based or physics-based atmospheric corrections. From a plethora of such corrections, which is best suited for vegetation and snow mapping in the mountainous Himalayan region needs to be decided. Hence, in this work, we evaluated a total of eight atmospheric corrections models including 5 image-based namely DOS (dark object subtraction method), improved dark object subtraction method (DOS3), COST (cosine theta), apparent reflectance (Aref), QUAC (QUick Atmospheric Correction), and 3 physics-based methods, namely SIAC (Sensor Invariant Atmospheric Correction), 6SV (Second Simulation of the Satellite Signal in the Solar Spectrum) and FLAASH (Fast Line-of-sight Atmospheric Analysis of Spectral Hypercubes). We found that 6SV and FLAASH were better than other methods and QUAC was the worst performer when applied to Landsat 8 OLI images of the Nepal Himalayan region which has dense vegetation and snow-covered areas. The better snow reflectance values were observed for FLAASH (B, G, R: 0.88, 0.89, 0.9; NIR: 0.83), SIAC (B, G, R: 0.85, 0.89, 0.89; NIR: 0.83) and 6SV (B, G, R: 0.87, 0.89, 0.89; NIR: 0.8) methods, whereas the FLAASH and SIAC methods exhibited higher vegetation reflectance values in the NIR band than other methods. The spectra from the standard spectral library were compared with the values of vegetation and snow spectral reflectance produced from corrected reflectance images. The mean values of snow and vegetation reflectance were higher for FLAASH, 6SV, and SIAC methods as compared to other methods. Therefore, FLAASH, 6SV, and SIAC methods, in contrast to other used atmospheric correction methods, have a high possibility of giving accurate snow and vegetation cover mapping. The snow cover and vegetation cover map prepared using NDSI and NDVI showed that areas covered under thin clouds and haze were better extracted when FLAASH, SIAC, and 6SV methods are applied as compared to other methods. Thus, this study confirms that physics-based atmospheric correction models such as FLAASH, SIAC, and 6SV methods should be used while working on satellite images of the Himalayan region where the focus is on snow and vegetation cover mapping.