The effect of different computed tomography (CT) scanning methods of the shoulder on image quality is uncertain. To compare the effect of different methods of CT scanning of the right shoulder on image quality and radiation dose. A total of 30 adults were divided into five groups. Group A received scans centered on the body's long axis, a scout direction of 0° + 90°, and automatic tube current modulation (ATCM). The other four groups (B, C, D, E) received isocenter scans centered on the shoulder with different scout directions (B and C: 0° + 90°, D: 0°, E: 0° + 270°) and tube currents (B: 420 mA; C, D, E: ATCM). The volume CT dose index (CTDIvol), dose-length product (DLP), image objective noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) were compared. Three subjective measures were also compared (noise, stripe artifacts, diagnostic confidence). The five groups differed significantly in all subjective and objective indexes. The CTDIvol and DLP decreased in the order of groups C, A, B, E, and D; the differences between groups A and B were not significant (P > 0.05). Groups B, C, and E had better SNR and CNR than groups A and D (P < 0.01). Subjective evaluations indicated group D was worse than groups B, C, and E (P < 0.05). In the ATCM system that uses the last scout view, CT of the shoulder should use isocenter scanning with the lateral scout view when the tube is away from the long axis of the body as the last execution direction.