Background Immobilization with shoulder braces is used for conservative treatment of an initial shoulder dislocation. Various arm positions have been investigated to determine optimal position in terms of the anatomical reduction of the Bankart lesion. Recently, the effect of immobilization in shoulder abduction as well as in external rotation has been reported. However, there are few studies assessing subjects' acceptability of the brace in various arm positions. We hypothesized that a certain arm position regarding abduction or external rotation for immobilization would induce significant discomfort during activities of daily living. Methods A dominant arm of 20 healthy participants was immobilized with a shoulder brace in 4 different positions; adduction and internal rotation (Add-IR), adduction and external rotation (Add-ER), 30° of abduction and 30° of external rotation (Abd-30ER), and 30° of abduction and 60° of external rotation (Abd-60ER). After completing immobilization for 24 h, subjects were asked to assess the discomfort of bracing for overall and individual activities, using a visual analogue scale. Data were compared among the four positions. Results For overall activities and several activities (eating, reading books, and removing and putting on pants), Abd-60ER was significantly more uncomfortable than adducted arm positions. Abd-30ER did not show any differences compared to the other arm positions. Conclusion Immobilization in abduction and external rotation seems to be acceptable although the arm position in 30° of abduction and 60° of external rotation is less comfortable than the others. Our results might be useful in determining and developing the ideal shoulder brace which could keep patients' compliance and improve their outcomes.