The aim of this study was to analyse quantitatively spine angle changes of subjects suffering from low back pain (LBP) during dynamic exercises. We explored the differences in the range of spine angle based on gender, disability severity, the correlation between the spine angle range and the visual analogue scale (VAS) scores, as well as the differences in standard deviations between the healthy and LBP subjects. We recruited thirty-nine LBP subjects and thirty-seven healthy people. They were asked to perform several movements from a standing position first and then from a sitting position. The motions were forward and backward bending, left and right lateral bending, as well as left and right axial rotation, respectively. Results show that for the most movements, the means of the spine angle changes in the females were larger than those in the males. In the LBP group, we observed much smaller spine angle values than those in the healthy subjects during exercise. With the increase of VAS score, a declining trend of the spine angle change was observed. There were significant differences in the spine angle range between standing and sitting positions when performing left and right axial rotation (p=0.000, p=0.002, respectively). We observed high correlations (with a max. result of r=0.804) for most movements, executed both from a standing and sitting position. We also found a wider range of standard deviation in the LBP subjects compared to healthy subjects. These results indicate that quantitative analysis of the spine angle range could provide an objective reference of the disability level, and allow for the progress assessment during the rehabilitation of low back pain patients.