Anorexia nervosa (AN) is an eating disorder of unknown origin that most commonly occurs in women and usually has its onset in adolescence. Patients with AN invariably have a disturbed body image and an intense fear of weight gain. There is currently no definitive treatment for this disease, which carries a 20% mortality over 20 years. Development of an appropriate animal model of AN has been difficult, as the etiology of this eating disorder likely involves a complex interaction between genetic, environmental, social, and cultural factors. In this review, we focus on several possible rodent models of AN. In our laboratory, we have developed and studied three different mouse models of AN based on clinical profiles of the disease; separation stress, activity, and diet restriction (DR). In addition, we discuss the spontaneous mouse mutation anx/anx and several mouse gene knockout models, which have resulted in an anorexic phenotype. We highlight what has been learned from each of these models and possibilities for future models. It is hoped that a combination of the study of such models, together with genetic and clinical studies in patients, will lead to more rational and successful prevention/treatment of this tragic, and often fatal, disease.