The purpose of this study was to examine the test-retest reliability of normalization methods for the infraspinatus muscle in a group of healthy subjects. Twelve healthy subjects (male=8, female=4) performed the maximal voluntary isometric contraction (MVIC) with examiner`s resistance, MVIC with a digital tension-meter (MVIC-DT), and sub-MVIC methods. Surface electromyography (EMG) signals were recorded from the infraspinatus muscles according to normalization methods. Reliability was analyzed using the intra-class coefficient (ICC), standard error of measurement (SEM), and minimal detectable difference (MDD). The results of the present study demonstrated that the sub-MVIC method has excellent test-retest reliability (ICC=0.92) with a relatively small SEM (5.9 mV) and MDD95 (16.4 mV), compared to MVIC-DT (ICC=0.73; SEM=11.2 mV; MDD95: 31 mV) and MVIC-E (ICC=0.5; SEM=15.7 mV; MDD95: 43.6 mV). These findings provide evidence that sub-MVIC is more appropriate for comparing the EMG activity for the infraspinatus muscle as a normalization method. If MVIC for normalization is needed, MVIC-DT is more appropriate than MVIC-E.