Nonsurgical treatment for carpal tunnel syndrome (CTS) is frequently offered to those with mild to moderate symptoms. However, the most sensitive electrophysiological parameters at follow-up, and most effective type of different methods of steroid treatment for CTS, remain unknown. The aim of this study was to compare the efficacy of local corticosteroid injection, phonophoresis, and iontophoresis for the treatment of CTS, and to evaluate the most sensitive electrophysiologic parameters at the follow-up visit. Different conservative treatments were applied in this prospective study. Fifty-six hands of 31 patients were randomly treated with local steroid injection, iontophoresis, or phonophoresis with 2- and 4-month follow-up visits. Conventional and new neurophysiologic sensorimotor nerve conduction parameters for CTS diagnosis were performed on the initial visit and at 2 and 4 months after treatment. The local steroid injection group showed a significant improvement in the mean nerve conduction parameters after the second and fourth months when compared with basal values, whereas no significant changes except the difference between the median and ulnar distal latency to the fourth digit (D4M-D4U), difference between sensory median distal latency to second digit and ulnar distal latency to the fifth digit (D2M-D5U), and median motor distal latency were observed in the iontophoresis and phonophoresis groups. At the end of follow-up period, the most sensitive parameters were D4M-D4U and D2M-D5U; the D4M-D4U was improved 60% in the injection group, 33% in iontophoresis group, and 39% in phonophoresis group, and the improvement of the D2M-D5U was 70%, 28%, and 28%, respectively, in the injection, iontophoresis, and phonophoresis groups. These findings show that steroid injection is superior to iontophoresis and phonophoresis in the treatment of CTS, and that the most sensitive neurophysiologic parameters in follow-up are D4M-D4U and D2M-D5U, being the objective measures of the outcome of CTS treatment.