Aim: To investigate the relation between the matching methods for different acupoints and nerve conduction velocity during the treatment of diabetic peripheral neuropathy with acupuncture. Methods: From January 2003 to May 2004, 100 patients, who accomplished courses they needed and were in accordance with the inclusion criteria, were selected from outpatients with diabetic peripheral neuropathy in the Department of Endocrinology, Xuanwu Hospital, Capital University of Medical Sciences. The 100 patients were divided into treatment groups of I, II, III and control according to their consultation time with 25 cases in each group. The method for treatment group I was acupuncturing into the first group of points including Dachui, Feishu, Kanshu, P'ishu, Weishu, Shenshu, xuanzhong(Chüehku), T'aihsi, Hsinchien and Janku, and into the second group of points including Kuanyüan, Chungfu, Ch'ime, Changmen, Chungwan, Chingmen, Ch'üch'ih, Hoku, Xuehai, Tsusanli and Sanyinjiao. The method for treatment group II was acupuncturing into the same points as in the first group of the treatment group I, and the choice of needles for acupuncture and acupuncture manipulation for treatment group II were same as those for treatment group I. The method for treatment group III was acupuncturing into the same points as in the second group of the treatment group I, and the manipulation was the same as for the treatment group II. The method for the control group was drug treatment. The nerve conduction velocities in all the groups were examined and recorded for comparison. Results: The motor nerve conduction velocities[(40.6 ± 3.4) m/s for median nerve and (37.0 ± 3.2) m/s for posterior tibial nerve] and sensory nerve conduction velocity[(46.4 ± 3.6) m/s for median nerve and (37.0 ± 3.9) m/s for posterior tibial nerve] in the treatment group I after treatment were significantly higher than those before treatment[(34.7 ± 2.9), (30.2 ± 3.0), (41.1 ± 3.1), (32.4 ± 3.3) m/s respectively] (t = 6.556, 7.753, 5.579, 4.51, P < 0.01). There was no significant difference in nerve conduction velocities in the control group before and after treatment(P > 0.05). After treatment, the motor and sensory nerve conduction velocities in the treatment group I were significantly higher those in the treatment groups II and III (t = 3.115 to 4.390, P < 0.01). Conclusion: The combination of the back-shu points and the front-mu points in the treatment of acupuncture can improve the nerve conduction velocity of patients with diabetic peripheral neuropathy.