Understanding lake ice growth and its sensitivity to climate change is vital to understand the thermal regime of thaw lake systems and predict their response to climate change. In this paper, a physically-based, two-dimensional, non-steady mathematical model is developed for studying the role of shallow tundra lakes in the Alaskan Arctic. Both the radiation absorption in lake water and the phase change in permafrost are considerd in the model. The materials the model includes are snow, ice, water, unfrozen and frozen soil (peat, silt, sand and gravel). The basic inputs to the model observed mean daily air temperature and snow depth. The ability of this model to simulate lake ice growth and thickness variation, lake water temperature distribution, the thermal regime of permafrost and talik dynamics beneath lakes, and thawing rate of permafrost below and adjacent to shallow thaw lakes offers the potential to describe the effects of climate change in the Alaskan Arctic.