The interactive multisensor snow and ice mapping system (IMS) of the National Oceanic and Atmospheric Administration combines multiple data sources to map northern hemisphere snow cover. IMS can identify snow cover beneath clouds using time series images from geostationary satellites and passive-microwave observations. During the snow disaster of 2008 in southern China, IMS snow-cover data were more accurate than those retrieved from passive-microwave remote sensing data and moderate resolution imaging spectroradiometer snow-cover products as compared with in situ measurements. The IMS snow-cover mapping accuracy was assessed against ground truth, which was derived using Landsat Enhanced Thematic Mapper Plus (ETM+) images. The actual snow cover was assessed from 47 ETM+ scenes that were obtained under mostly clear-sky conditions (cloud cover of <20%) from 2008 to 2011 and were subsequently used to evaluate the IMS snow-cover product. Land cover and terrain effects on the accuracy of snow-cover products were considered in this study. The IMS snow-cover product was consistent with the ETM+ snow images over flat surfaces, e.g., cropland, and the average agreement was greater than 85%. For forested, mountainous areas, a pronounced inconsistency was observed between the two datasets. The agreement of the IMS snow-cover product in these regions was <75%, and the IMS appeared to overestimate snow by over 50%. The sparser snow in 2009, 2010, and 2011 caused poorer accuracies and more severe overestimations. In addition, mixed pixels, particularly in complex terrain, have been recognized as a major problem that affects the accuracy of IMS snow detection because of the product’s coarse spatial resolution (i.e., 4 km). Specifically, fragmented snow cover is difficult to discern with 4-km pixels. Therefore, further studies are required to develop a fractional snow-cover algorithm for the IMS product.