Glycogen phosphorylase is an important therapeutic target for the potential treatment of type 2 diabetes. The importance of computation in the search for potent, selective and drug-like glycogen phosphorylase inhibitors which may eventually lead to antihyperglycemic drugs is now firmly established. Acting solo or more effectively in combination with experiment in a multidisciplinary approach to structure based drug design, current day modeling methods are an effective means of reducing the time and money spent on costly experimental procedures. Glycogen phosphorylase is an allosteric protein with five different ligand binding sites, hence offering multiple opportunities for modulation of enzyme activity. However, the binding sites have their own individual characteristics, so that different modeling approaches may be more effective for each. This review is focused on advances in the modelling and design of new inhibitors of the enzyme aimed towards providing the reader with some useful hints towards more successful computer-aided inhibitor (drug) design targeting glycogen phosphorylase.