Ｏｂｊｅｃｔｉｖｅ　To observe the effects of preemptive analgesia with continuous femoral nerve block（CFNB） after total knee arthroplasty（TKA） in elderly patients on postoperative pain scores and early postoperative cognitive function.　Methods　Sixty patients aged more than 65 years with American society of anesthesiologists（ASA） physical status Ⅰ-Ⅱ， who were scheduled for elective total knee arthroplasty， were randomly devided into 2 groups（n=30）： preemptive analgesia group and control group. Preemptive analgesia group was implemented CFNB before surgery， while control group was done CFNB after surgery. Patient's general information before operation was recorded. The changes of hemodynamics and the value of bispectral index（BIS） were monitored during operation. The doses of all anesthetics， the time of operation， tourniquet and anesthesia were also recorded. Cognitive function was measured on 1 d before surgery， 1， 3， 5 d and 7 d after surgery using mini-mental state examination（MMSE）. The occurrence of postoperative cognitive dysfunction（POCD） was defined when the decreased MMSE score after surgery was more than one standard deviation. The visual analog scale（VAS） at the same days in resting and exercise were recorded to evaluate the degree of postoperative pain.　Results　The total dosage of remifentanil in preemptive analgesia group （1.7±0.4） mg was less than that in control group（1.9±0.3） mg（P<0.05）. VAS exercise score in preemptive analgesia group were lower than that of control group（P<0.01）. VAS resting scores in two groups had no significant significance（P>0.05）. The incidence of POCD in preemptive analgesia group（6.7%） was lower than in control group（26.7%）（P<0.05）.　Conclusions　The preemptive analgesia produced by preoperative CFNB can improve the recovery of early postoperative cognitive function in elderly patients performed TKA. This can be related to the decreasing of the degree of postoperative exercise pain.