Objectives Determining gait disorders in boys with Duchenne's Muscular Dystrophy (DMD) through a longitudinal study. Material and method Retrospective study of 19 gait analyses recorded from nine boys with confirmed DMD according to the same protocol. They were 6 years 1 month to 10 years 5 months old at the first record. Clinical examination included : anthropometric data (size, weight), muscle strength assessment in similar conditions and distanced from any physical effort, according to the MRC Grading of muscle strength (MRCG). Gait analysis was performed by a VICON ® system, with five infra-red videocameras, two force-plates and a electromyography recording set. The kid walked barefoot, without aid, along a 10m long track. Results Gait parameters were compared to those of boys who were similar concerning age and anthropometric data. Muscular strength decrease was correlated to age (R = −0.54 ; p < 0.02). Stride length decreased, with an initial increase of cadence in order to maintain speed, then cadence and speed dropped when the disease became worse, with an inverse correlation between cadence and MRCG score (R = −0.46 ; p < 0.05) and between age and cadence (R = −0.46 ; p < 0.05). Lumbar hyperlordosis and pelvic tilt increased with age (R = 0.70, p < 0.01) and were correlated to hip extension decrease at standing phase (R = −0.78 ; p < 0.01) and also to hip and knee flexion increase at swing phase (R = 0.6 p < 0.01 et R = 0.51 p < 0.02 respectively). At initial contact, ankle was plantar flexed, and plantar flexion range was correlated to pelvic tilt (R = 0,45, p < 0.05). Ground reaction force (GRF) relations with centres of ankle, knee and hip in sagittal plane were particularly relevant : — at initial contact, GRF was pointed forwards in front of the knee and hip centre ; — at standing phase, it became vertical and stayed in front of knee centre, but behind hip centre ; — at toe-off, it intersected knee centre, while staying behind hip ; — during double support, it was at that time behind the knee. In order to determine this GRF vector direction during the whole gait cycle, the ankle needed to be always in equinus position and the centre of pressure was always in front of the heel. Conclusion Gait of DMD boys was initially compensatory of gluteus insufficiency and its maintain seemed directly linked to active equinus as long as triceps surae strength was sufficient, generation power of ankle at toe-off being not really different from healthy boys of the same age, and until advanced stage of the disease.