Radiographic Cobb's angle is the gold standard for evaluation of spinal curvature, however X-ray is ionizing. In contrast, ultrasound is non-ionizing and inexpensive, thus more accessible. However, no study has reported the reliability and accuracy of ultrasound on sagittal curvature analysis. Ultrasound and X-ray scanning were conducted on 16 sets of spine phantoms with different deformities. Intrarater and inter-rater reliability, correlations, mean absolute differences (MAD) and linear regression of Ultrasound spinous process angles (USSPA), X-ray spinous process angles (XSPA) and X-ray Cobb's angles (XCA) together with the intra-operator reliability of USSPA were investigated. In addition, USSPA and XCA of five AIS subjects were scanned using the ultrasound system. In the phantom study, excellent intra-rater and inter-rater reproducibility for the three angles and excellent intraoperator reproducibility for USSPA were demonstrated. Good to moderate correlations were obtained between lumbar XCA and XSPA and between lumbar XCA and USSPA, whereas excellent correlations were observed between the other angles. All three angles indicated positive linear relationships, with MAD of 5.8°, 3.0° and 6.0° for XSPA against XCA, USSPA against XSPA and USSPA against XCA, respectively. The results of the preliminary study demonstrated a high intrareliability for the ultrasound measurements. The measured difference between the USSPA and XCA methods was 6.3° ± 5.4°. It should be noted that ultrasound and X-ray measurements were based on different structures of the vertebrae. The results showed that ultrasound is feasible for measuring sagittal curvature and has the potential for monitoring the curve progression and evaluating sagittal spinal profiles.