The majority of candidates for epilepsy surgery, in both children and adults, belong to the syndromic category of partial epilepsies. However, particularly in children, the clinical expression of epilepsy may sometimes be misleading, as paroxysmal events may present as generalized seizures although having a focal onset. The spectrum of surgical possibilities for early-onset epilepsy has tended to widen rapidly and to include not only other focal epilepsies but also more difficult types. This is because developmental lesions that are the major cause of infantile epilepsy are often poorly localized, involve extensive brain areas and require extensive operations (e.g. hemispherotomy). Surgery is also used for progressive conditions such as Rasmussen’s encephalitis or Sturge-Weber syndrome. Some forms of generalized symptomatic epilepsy may also benefit from palliative surgical procedures (callosotomy in cases with repetitive drop attacks often as a manifestation of a Lennox-Gastaut syndrome). The timing of surgery partly depends on the probable effectiveness of the operation available for each particular patient. Candidates for epilepsy surgery must be identified early in the process of the disease, thus increasing the chances for a satisfactory cognitive and behavioral outcome. The decision should be taken by experienced multidisciplinary groups, that will also ensure post-operative follow-up, both in terms of medical management and psycho-social integration. Prospective studies are still needed, to evaluate the long-term cognitive evolution of children operated on early for their epilepsy and cured. Increasing evidence indicates that drug therapy is likely to fail to achieve control of the seizures when two-three of the major drugs, properly chosen by competent clinicians, have not obtained satisfactory results, i.e. full control of seizures and absence of side effects. Testing of all possible drugs is not advisable, because this process would be excessively long when surgery is a reasonable possibility. The requirements for resective surgery in children with partial drug-resistant epilepsy vary with the type of resection considered. Three basic requirements apply to almost all cases: (a) the epileptogenic area must be localized to a territory whose removal is contemplated; (b) that no other independent epileptogenic area exists in those areas that are not included in the planned resection, and (c) that any possible deficit resulting from resection must be acceptable. The current experience confirms that the complete resection of the epileptogenic area is the major condition for a satisfactory surgical result. The possibilities for resective surgery depend upon the localization of the ictal onset zone, its relationship with adjacent functional brain areas and the availability of convergent data that point to a single localization. Epilepsy surgery mandates a multidisciplinary approach that requires special skills and sophisticated instruments and materials that cannot be improvised. A global evaluation of risks and expected benefits is always required. All decisions require a close collaboration between the epilepsy team, the patient and the family. With the exception of high quality MRI and video-EEG recording of seizures, the need to perform any other complementary presurgical investigation should be evaluated on an individual basis.