Abstract This article describes an attempt to map snow cover accurately from other land covers using Moderate Resolution Imaging Spectrometer (MODIS) data of 500 m spatial resolution. The workflow includes reflectance modelling, computing snow-cover fraction (SCF) and establishing an empirical relationship between the SCF and normalized difference snow index (NDSI) to map snow cover at operational level. Regression relationships have been developed between the SCF derived from the linear mixture model (LMM) and snow obtained from the NDSI based on two criteria, namely: SCF greater than 0.0 and SCF greater than 0.1. The best regression equation has been selected by examining respective graph plots using statistical measures of mean absolute error, correlation coefficient, root mean square error (RMSE) and uncertainty analysis. The results have been validated against the actual SCF obtained from a high-resolution 15 m Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) visible and near infrared (VNIR) scene and covering a substantial range of snow cover of the same area. The selected regression model SCF = 0.25 + 0.35 × NDSI has been tested on other areas and validation efforts show that the pixel-level SCF relationship provides useful results as measured in independent tests against actual SCF obtained from ASTER scene.