To compare the intra- and interobserver reliability of three-dimensional (3D) volumetric versus conventional two-dimensional (2D) power Doppler ultrasonography (US) in the assessment of peripheral enthesitis in spondylarthritis (SpA). Sixteen patients with SpA according to ASAS criteria were included. Two rheumatologists (one experimented in musculoskeletal US (sonographer 1) and one beginner (sonographer 2)) performed independently a 2D US scoring of the enthesis using the Madrid Sonographic Enthesis Index score followed by a 3D acquisition of the same entheseal sites. The reading of the 3D acquisition was performed a minimum of 1 week apart. Intraobserver reliability was evaluated by a second reading of the same images. The duration of 2D US scanning, 3D US acquisition and reading was recorded. Intraclass correlation coefficients (ICCs) were used for the reliability analysis. Intraobserver reproducibility was good to excellent for 2D US and good for 3D US (ICC (95 %CI) 2D US 0.776 (0.471-0.916) and 0.96 (0.892-0.986) and ICC (95 %CI) 3D US 0.796 (0.498-0.921) and 0.703 (0.325-0.886) for sonographer 1 and 2, respectively). Interobserver reliability was slightly better for 3D US than for 2D US (ICC (95 %CI) 0.776 (0.471-0.916) for 3D US versus 0.641 (0.221-0.859) for 2D US). The mean time (±SD) for 2D US scanning was 23 min (±4) whereas the mean time for 3D US volume acquisition and reading was 16.5 min (±2.6) (p < 0.001). 3D US showed good intra- and interobserver reliability in the assessment of enthesitis in SpA and shortened the needed time for scanning. It can be performed by a nonexperienced examiner without loss of reliability.