Objective To investigate the different memory outcomes in temporal lobe epilepsy patients underwent different surgical approaches.Methods Two hundred forty-eight patients with temporal lobe epilepsy and hippocampal scle-rosis underwent standard anterior temporal lobectomy ( ATL, n=83 ) or selective amygdalohippocampectomy ( SAH, n=165) from 2009 to 2013.All the patients underwent clinical memory function assessment before surgery, 3 months and 2 years after surgery respectively.Results The memory quotient ( MQ) of patients who underwent brain surgery in the domi-nant hemisphere significantly decreased 3 months after surgery (74.5 ±16.2, 75.6 ±19.5) compared to presurgery MQ (82.9 ±15.8, 83.2 ±21.2) in both ATL and SAH groups (P<0.05).Although MQ was slightly recovered at 2 years af-ter surgery, MQ (75.1 ±14.1, 76.1 ±17.6) was still significantly lower compared with presurgery MQ (P<0.05).A-mong this, both the decrease extent of the MQ 3 months after surgery and 2 years after surgery were smaller in the SAH group than in the ATL group (7.6 vs.8.4;7.1 vs.7.8).The MQ of patients who underwent brain surgery in the non-dominant hemisphere (either ATL or SAH ) increased slightly 3 months after surgery (87.2 ±15.1, 88.1 ±16.9) com-pared to presurgery MQ (85.5 ±13.5, 85.3 ±19.7) although the difference was not statistically significant.The MQ of these two groups improved significantly 2 years after surgery (92.8 ±12.7, 93.7 ±17.1)(P<0.05).The improvement extent of the MQ was larger in the SAH group than in the ATL group (8.4 vs.7.3).Conclusions SAH may be better than ATL in the maintenance of memory function in patients with temporal lobe epilepsy and hippocampal sclerosis.