AbstractCrystalline polysaccharides are abundant in nature and can be transformed into highly functional materials. However, the molecular basis for the formation of higher-order structures remains unclear. Computer simulation is an advanced tool for modeling macromolecular structures, and the atomistic simulations provide valuable information on the crystalline polysaccharides. Fiber deformation, crystalline transition, and novel nanostructures of cellulose were characterized through molecular dynamics simulations and density functional theory calculations of models of molecular chain sheets extracted from the crystal structure of the cellulose polymorphs. Extended ensemble molecular dynamics simulations were applied to analyze the artificial crystal structure of non-natural amylose analog polysaccharides, revealing the hexagonal packing of double helices through the self-assembly of molecular chains dispersed in aqueous solution. Dissolution simulations of the cellulose and chitin crystalline fibers revealed that the anions of ionic liquids, with their solvation power, played a key role in the cleavage of intermolecular hydrogen bonds in the crystal structure, whereas the cations contributed to irreversible molecular chain dispersion. The good correlation between the actual solubility of polysaccharides and the predicted number of intermolecular hydrogen bonds prompted the development of a platform that combined simulations and machine learning for high-throughput screening of solvents for cellulose and chitin.