Background Lumbar spondylolysis is a prevalent cause of low back pain. Our research group previously proposed the MT classification based on CT pars defect width: Type I (appositional type, defect < 1 mm) and Type II (separated type, defect ≥ 2 mm). Static morphological differences between the two subtypes have been verified, but dynamic functional gait differences remain unclear.Methods This cross-sectional study enrolled 96 participants matched by age, sex, BMI, and height, including 32 Type I patients, 32 Type II patients, and 32 healthy controls. Gait data were collected using a markerless computer vision–based motion capture system, and spatiotemporal parameters, lower-limb kinematics, and lumbar three-dimensional motion were analyzed.Results No significant intergroup differences were observed in baseline characteristics. Compared with Type I patients and controls, Type II patients exhibited significantly slower gait speed, shorter step length, higher stance phase proportion, and lower swing phase proportion. Type II patients also showed reduced hip flexion, increased knee extension, increased lumbar lateral bending and extension, and decreased lumbar flexion and rotation. Gait patterns of Type I patients were similar to those of healthy controls.Conclusions Type II lumbar spondylolysis is associated with more severe gait dysfunction and functional impairment. Computer vision–based motion capture gait analysis can objectively reflect biomechanical differences between Type I and Type II subtypes, providing dynamic evidence for clinical classification, assessment, and individualized treatment.