Molecular recognition and binding are two very important processes in virtually all biological and chemical processes. An extremely interesting system involving recognition and binding is that of thermal hysteresis proteins at the ice-water interface. These proteins are of great scientific interest because of their antifreeze activity. Certain fish, insects and plants living in cold weather regions are known to generate these proteins for survival. A detailed molecular understanding of how these proteins work could assist in developing synthetic analogs for use in industry. Although the shapes of these proteins vary from completely alpha-helical to globular, they perform the same function. It is the shapes of these proteins that control their recognition and binding to a specific face of ice. Thermal hysteresis proteins modify the morphology of the ice crystal, thereby depressing the freezing point. Currently there are three hypotheses proposed with respect to the antifreeze activity of thermal hysteresis proteins. From structure-function experiments, ice etching experiments, X-ray structures and computer modeling at the ice-vacuum interface, the first recognition and binding hypothesis was proposed and stated that a lattice match of the ice oxygens with hydrogen-bonding groups on the proteins was important. Additional mutagenesis experiments and computer simulations have lead to the second hypothesis, which asserted that the hydrophobic portion of the amphiphilic helix of the type I thermal hysteresis proteins accumulates at the ice-water interface. A third hypothesis, also based on mutagenesis experiments and computer simulations, suggests that the thermal hysteresis proteins accumulate in the ice-water interface and actually influence the specific ice plane to which the thermal hysteresis protein ultimately binds. The first two hypotheses emphasize the aspect of the protein 'binding or accumulating' to specific faces of ice, while the third suggests that the protein assists in the development of the binding site. Our modeling and analysis supports the third hypothesis, however, the first two cannot be completely ruled out at this time. The objective of this paper is to review the computational and experimental efforts during the past 20 years to elucidate the recognition and binding of thermal hysteresis proteins at the ice-vacuum and ice-water interface.