A method for mapping snow cover using Advanced Very High Resolution Radiometer (AVHRR)/Normalized Difference Vegetation Index (NDVI) 10-day composite data was developed at the continental scale. The 10-day AVHRR/NDVI composite data and snow depth data of AmeDas (Automated Meteorological Data Acquisition System) in l999 were compared to extract the snow cover land surface by determining NDVI threshold values. Initially, one threshold value was applied to all overJapan. However, the results showed inaccuracies in snow map for the different land cover effects the NDVI values greatly. To improve the estimation accuracy, classifying the land cover into 5 categories was conducted. The snow maps derived from different threshold values for different land cover categories instead of one threshold value for all over Japan showed better accuracy. The areas where the accuracy of mapping got better are in the forest area such as alpine vegetation and urban area especially around Kanto.