Water and ice interfaces are of crucial importance in our modern life, but our knowledge about them are still at a primitive stage. We describe in this talk our attempt of using surface-specific sum-frequency vibrational spectroscopy to search for molecular-level understanding of the interfaces.