Accurate monitoring of the spatiotemporal dynamics of snow and ice is essential for under-standing and predicting the impacts of climate change on Arctic ecosystems and their feedback on global climate. Traditional optical and Synthetic Aperture Radar (SAR) remote sensing still have limitations in the long-time series observation of polar regions. Although several studies have demonstrated the potential of moonlight remote sensing for mapping polar snow/ice covers, systematic evaluation on applying moonlight remote sensing to monitoring spatiotemporal dynamics of polar snow/ice covers, especially during polar night periods is highly demanded. Here we present a systematic assessment in Svalbard, Norway and using data taken from the Visible Infrared Imaging Radiometer Suite (VIIRS) on the Suomi National Polar-orbiting Partnership (SNPP) Day/Night Band (DNB) sensor to monitor the spatiotemporal dynamics of snow/ice covers during dark Arctic winters when no solar illumination available for months. We successfully revealed the spatiotemporal dynamics of snow/ice covers from 2012 to 2022 during polar night/winter periods, using the VIIRS/DNB time series data and the object-oriented Random Forests (RF) algorithm, achieving the average accuracy and kappa coefficient of 96.27% and 0.93, respectively. Our findings indicate that the polar snow/ice covers show seasonal and inter-seasonal dynamics, thus requiring more frequent observations. Our results confirm and realize the potential of moonlight remote sensing for continuous monitoring of snow/ice in the Arctic region and together with other types of remote sensing data, moonlight remote sensing will be a very useful tool for polar studies and climate change.