Important developments have taken place in cognitive theory of eating disorders (EDs) (and also in other disorders) since the review paper published by M.J. Cooper in 1997. The relevant empirical database has also expanded. Nevertheless, cognitive therapy for anorexia nervosa and bulimia nervosa, although helpful to many patients, leaves much to be desired. The current paper reviews the relevant empirical evidence collected, and the theoretical revisions that have been made to cognitive models of eating disorders, since 1997. The status and limitations of these developments are considered, including whether or not they meet the criteria for "good" theory. New theoretical developments relevant to cognitive explanations of eating disorders (second generation theories) are then presented, and the preliminary evidence that supports these is briefly reviewed. The lack of integration between cognitive theories of EDs and risk (vulnerability) factor research is noted, and a potential model that unites the two is noted. The implications of the review for future research and the development of cognitive theory in eating disorders are then discussed. These include the need for study of cognitive constructs not yet fully integrated (or indeed not yet applied clinically) into current theories and the need for cognitive theories of eating disorders to continue to evolve (as they have indeed done since 1997) in order to fully integrate such constructs. Treatment studies incorporating these new developments also urgently need to be undertaken.