Scoliosis is a spinal disorder that is highly prevalent in the adolescent population. To address the radiation hazards associated with X-ray assessment of scoliosis, this article designs and implements a personalized modeling and scoliosis assessment system for the spine based on morphological recognition. First, the feature points related to the spine are extracted, and the additional feature points are generated according to the relative positional relationship between the feature points. Secondly, the feature point correction algorithm and the filtering algorithm are designed and applied to improve the positional accuracy of the feature points. Finally, the vertebral bone model is aligned to the spine line fitted by the feature points in Unity to obtain a personalized 3D spine model, and the Cobb angle, the thoracic kyphosis angle, and the lumbar lordosis to assess scoliosis. An experiment is implemented on 28 subjects to compare and analyze the results of the systematic assessment with those of the X-ray assessment. The Pearson correlation coefficient between the Cobb angle and the actual Cobb angle is 0. 82, with a mean absolute error of 3. 4° and a root-mean-square error of 4. 2°. The Pearson correlation coefficient between the thoracic kyphosis angle and the actual thoracic kyphosis angle is 0. 80, with a mean absolute error of 3. 4° and a root-mean-square error of 3. 8°. The Pearson correlation coefficient between the lumbar lordosis and the actual lumbar lordosis is 0. 78, the mean absolute error is 3. 2°, and the root mean square error is 3. 7°. The experimental results show that the scoliosis assessment system is highly accurate, which is easy to use. It can be applied to scoliosis screening in a large adolescent population.