Purpose With the advent of new very selective techniques like thermal laser ablation to treat drug-resistant focal epilepsy, the controversy of resection size in relation to seizure outcome versus cognitive deficits has gained new relevance. The purpose of this study was to test the influence of the selective amygdalohippocampectomy (SAH) versus nonselective temporal lobe resection (TLR) on seizure outcome and cognition in patients with mesial temporal lobe epilepsy (MTLE) and histopathological verified hippocampal sclerosis (HS). Methods We identified 108 adults (> 16 years) with HS, operated between 1995 and 2009 in Denmark. Exclusion criteria are the following: Intelligence below normal range, right hemisphere dominance, other native languages than Danish, dual pathology, and missing follow-up data. Thus, 56 patients were analyzed. The patients were allocated to SAH (n = 22) or TLR (n = 34) based on intraoperative electrocorticography. Verbal learning and verbal memory were tested pre- and postsurgery. Results Seizure outcome did not differ between patients operated using the SAH versus the TLR at 1 year (p = 0.951) nor at 7 years (p = 0.177). Verbal learning was more affected in patients resected in the left hemisphere than in the right (p = 0.002). In patients with left-sided TLR, a worsening in verbal memory performance was found (p = 0.011). Altogether, 73% were seizure-free for 1 year and 64% for 7 years after surgery. Conclusion In patients with drug-resistant focal MTLE, HS and no magnetic resonance imaging (MRI) signs of dual pathology, selective amygdalohippocampectomy results in sustained seizure freedom and better memory function compared with patients operated with nonselective temporal lobe resection.