Natural lakes play a crucial role in the Earth's water storage, holding significantly more water than all existing and planned reservoirs. However, their importance is often underestimated in global hydrological and human-Earth system models, with crucial aspects such as surface area dynamics, socio-economic impacts of smaller lakes, and surface phenomena like snow depth, snow cover, and ice formation overlooked. This oversight results in a considerable knowledge gap in our understanding of the Earth's water cycle and the potential implications of climate change for multi-sector dynamics. To address this issue, we have developed Xanthos-Lake, a novel global lake water balance model that accounts for over 1.4 million lakes documented in HydroLAKES dataset. Xanthos-Lake remaps HydroLAKES onto 0.5-degree grids and differentiates three categories of lakes within each grid based on their contributing area (lake area + drainage area). It then merges these categories to form small, medium, and large lakes, which are arranged in a cascade manner. The bathymetry of these merged lakes is regenerated using data from the GLOBathy dataset. Moreover, surface processes like lake ice thickness, snow cover, and snow depth are modeled using deep learning algorithms (e.g., LSTM). Integrated with Xanthos, a global hydrological model designed for use with Global Change Analysis Model (GCAM), Xanthos-Lake enhances our comprehension of global freshwater supplies. Preliminary testing indicates that the new model captures the monthly surface area dynamics, lake evaporation, ice thickness, snow depth, and snow cover reasonably well. This suggests that Xanthos-Lake can provide valuable insights for studying the Earth's water cycle and exploring the potential impacts of climate change on various sectors. The integration facilitates a more comprehensive evaluation of the role of various freshwater sources in shaping the intricate interactions and interdependencies of human and natural systems across scales.