This study evaluated the agreement between Kinect V2 with AssessLink and Vicon motion capture systems in children performing the vertical jump (JL) and forward leap (Leap). Eighteen healthy children (5-17 years) performed the movements while recorded simultaneously by both motion capture systems. Spearman correlation ([rho]) and Bland-Altman were used to determine joint angle agreement. Overall, agreement varied by body segment, movement type, and phase. JL for knees and shoulder ranged from [rho] = -0.220 to 0.543; bias = 8.1 to 23.0° and [rho] = -0.743 to 0.410; bias = 8.1 to 56.5°, respectively. Leap had poorer correlation and bias for the appendicular segments ([rho] = -0.091 to 0.306; bias = -47.5 to 73.7°) than axial segments ([rho] = 0.345 to 0.553; bias = 11.3 to 22.9°). The Kinect V2 should not replace criterion kinematic systems, but may improve gross motor assessment methods when selecting for specific movement observations.