Objective: To quantify gait abnormalities in people with Dravet syndrome (DS).Methods: Individuals with a confirmed diagnosis of DS were enrolled. They were stratified according to knee flexion at initial contact (IC) and range of motion (ROM) during stance [crouch gait if knee flexion >20° at IC and knee range of motion (ROM) >15°; straight gait: knee flexion <20° at IC]. A 1D ANOVA (α = 0.05) was used to test statistical differences among the joint kinematics and spatio-temporal parameters of the cohort and an age-matched control group. Clinical (neurological and orthopedic evaluation) and anamnestic data (seizure type, drugs, genetic mutation) were collected; distribution between the two gait phenotypes was assessed with the Fisher exact test and, for mutation, with the chi-squared test (p-value < 0.05).Results: Seventy-one subjects were enrolled and evaluated with instrumented gait analysis. Fifty-two were included in the final analysis (mean age 13.8 ± 7.3; M 26). Two gait patterns were detected: an atypical crouch gait (34.6%) with increased ankle, knee and hip flexion during stance, and reduced walking speed and stride length without any muscle-tendon retraction; and a pattern resembling those of healthy age-matched controls, but still showing reduced walking speed and stride length. No difference in clinical or anamnestic data emerged between the two groups.Significance: Objectively quantified gait in DS shows two gait patterns with no clear-cut relation to clinical data. Kinematic parameters abnormalities are likely due to stabilization issues. These findings may guide rehabilitative and preventive measures.