Surface electromyogram (sEMG) is one of the evaluation method of muscle function objectively and noninvasively. In recording sEMG, it need to avoid innervation zone (IZ). However, it can’t find on a skin surface, and a relative position between the electrodes and IZ position changes in changing the muscle length. Researchers usually estimate IZ position from signals measured by visually. This is inferior in objectivity, and it always need to estimate by a researcher. The purpose of this study was to develop algorithm of estimating innervation zone automatically. In this algorithm, the characteristic waves were extracted, and IZ location was estimated using cross-correlation and amplitude ratio between the neighboring channels. To consider efficacy of this algorithm, we carried out experiments to compare IZ estimation in visual and this algorithm during dynamic contraction. As the result, the rate of concordance between visual observation and this algorithm was over 90 percent. Therefore, it is considering that the algorithm is effective measures of IZ estimation.