Quantitative data assessment on the basis of three-dimensional gait analysis has been routinely used in the evaluation of pathological gait of children with cerebral palsy. However, a similar quantitative methodology has not been applied for spina bifida patients in whom atypical gait patterns are thought to correlate with various levels of neurological paralysis. The purpose of this study is to investigate the differences among gait patterns in spina bifida between different levels of neurological lesions using quantitative methods: Gait profile score (GPS) and gait variable scores (GVS), scoring subject's gait deviation from a reference. In this cross-sectional study, 22 children with spina bifida (11 women, 11 men; mean age 9.4 years, SD 3.8 years, range 3-17 years), were examined using three-dimensional gait analysis from 2008 to 2018. Physical examination allowed for classification of each of the 44 limbs as either L4, L5 or S1 and comparison with the GPS and GVS using a linear mixed model. GPS and the GVS of the pelvis and hip range of motion in the coronal plane were significantly higher in the L4 group than in the L5 and S1 groups (GPS, P = 0.041, P = 0.003, respectively; GVS of pelvis, P = 0.001, P = 0.001; GVS of hip, P < 0.001, P < 0.001) GVS (foot progression angle) was significantly lower in the S1 group than in L4 and L5 groups (P < 0.001, P = 0.037). We found that GPS and GVS enable us to quantitatively assess the differences among gait patterns between different neurological levels. The scoring tool showed the potential for detecting individual neurological changes.