The purposes ofthis study were to calculate and verify the center of rotation (COR) using least squares method and to determine the displacement of the humeral head during passive abduction movements of the glenohumeral joint. A hinge and pivot joint model with a reflective marker triad rigidly attached was used to verify the accuracy of the COR estimation. Eleven fresh cadaver shoulder specimens with reflective markers attached to the humerus were used to find the COR of the glenohumeral joint. A 6-camera motion analysis system (Vicon370) was used to monitor the trajectories of these reflective markers. In the hinge joint model,the errors between the estimated COR and the real COR were 0.25， 0.13 and 0.88 mm in the x-，y- and z-directions， respectively. In the cadaver model， the means of standard deviation of the estimated COR in the X-， y- and z-directions were 1.16，0.72and 1.12mm， respectively. Since reflective markers were attached rigidly on the humerus， the error caused by the skin movements could be ignored. The results showed that the least squares method is suitable for determining the COR of the glenohumeral joint in vitro