Twenty to 30% of patients with epilepsy develop chronic or intractable epilepsy, i.e. the seizures persist despite accurate diagnosis and carefully monitored treatment with antiepileptic drugs. It is still not known why and how epilepsy becomes an intractable disorder in such patients, while other patients with seemingly identical seizure types can achieve control of seizures with medication. Experimental epilepsy research has generated new neurophysiological, neurochemical and neuropharmacological approaches but is still hampered by a lack of adequate experimental models of chronic intractable epilepsy. An animal model of epilepsy allowing selection of pharmacoresistant and pharmacosensitive subgroups of animals would be a valuable tool to study mechanisms of intractability and to develop more effective treatment strategies. This review concentrates on identifying animal models that mimic patterns of pharmacological resistance in humans with epilepsy. Two models seem to be interesting in this regard, epileptic dogs with different types of spontaneous recurrent seizures and amygdala-kindled rats. In both models, animals which do not respond to repeated or chronic administration of antiepileptic drugs (non-responders) can be separated from animals in whom antiepileptics are effective (responders). Unfortunately, the dog model has several inherent logistical problems, whereas pharmacoresistant subgroups of kindled rats offer a unique tool to study why seizures become intractable, particularly because pathophysiologically processes in these resistant rats can be compared directly with those of kindled rats which respond to treatment. Furthermore, the new rat model can be used to identify predictors of intractability during onset of treatment with an antiepileptic drug and is a valuable addendum to the battery of animal models used in preclinical evaluation of novel drugs. In addition to pharmacoresistant epileptic dogs and kindled rats, several other potential models of intractable epilepsy are described to stimulate and outline areas for future research.