Objective To assess the reproducibility of shear wave elastography (SWE) in measuring the stiffness of the frozen shoulder coracohumeral ligament in different shoulder positions. Methods Two experienced sonographers conducted SWE on the coracohumeral ligament to measure the mean elastic modulus, representing its stiffness, in 45 patients presenting with stages Ⅱ and Ⅲ of frozen shoulder. The second measurement was carried out one week following the initial assessment. SWE measurements were taken in the neutral shoulder position, as well as at 10°, 20°, and 30° of external rotation. The intra- and inter-examiner reproducibility of the coracohumeral ligament"s elasticity modulus in these four shoulder positions was assessed using a two-way randomized model with an intragroup correlation coefficient (ICC). The normal distribution of inter- and intra-examiner differences in measurements was tested using the Shapiro-Wilk test. Differences that followed a normal distribution were visualized on Bland-Altman plots to determine the 95% confidence interval (CI) limits of agreement for mean differences between and within examiners. Results There was a statistically significant difference in coracohumeral ligament stiffness measured within and between examiner groups when the shoulder was in neutral position, 10° and 20° of external rotation (P < 0.05); the reproducibility of coracohumeral ligament stiffness measured within and between examiner groups was poor, and the ICCs were both < 0.40. When the shoulder joint was externally rotated at 30°, the differences in coracohumeral ligament stiffness measured within and between groups were not statistically significant (P<0.05), and the reproducibility of coracohumeral ligament stiffness measured within and between groups was good: the ICC was 0.645 (95% CI 0.439~0.788, P=0.000) for examiner A, 0.619 (95% CI 0.399~0.771, P=0.000) for examiner B; and the ICC at first measurement between two examiners was 0.659 (95% CI 0.458~0.771). (95% CI 0.399 to 0.771, P=0.000); between the two examiner groups the ICC was 0.659 (95% CI 0.458 to 0.797, P=0.000) for the 1st measurement and 0.543 (95% CI 0.297 to 0.721, P=0.000) for the 2nd measurement. The initial ICC measurement between the two examiners was 0.659 (95% CI 0.458-0.797, P=0.000), and the subsequent ICC measurement was 0.543 (95% CI 0.297-0.721, P=0.000). Furthermore, the Shapiro-Wilk test for differences in intra- and intergroup measurements of the examiner"s group at 30° of external rotation of the shoulder follows a normal distribution. The Bland-Altman scatterplot suggested that at 30° of shoulder external rotation, 98% of the points within the 95% limits of agreement (LOA) fell within the 95% LOA for the intragroup measurements in Examiner Group A, and 93% for the intragroup measurements in Examiner Group B. The number of points that fell within the 95% LOA for the first between-group measurements was 96%, and the number of points that fell within the 95% LOA for the second between-group measurements was 98%; the consistency between the intra- and between-group measurements was good. Conclusion The SWE measures stiffness of the coracohumerale ligament in patients with stage Ⅱ and Ⅲ frozen shoulder, and values measured using 30° of external rotation of the shoulder joint have good repeatability and checking consistency.