This work presents a novel experimental method for performing dynamic yarn pull-out using a modified tensile split Hopkinson pressure bar. The pull yarn is fixed to a piezoelectric load cell, while the woven fabric specimen is displaced at pull-out velocities ranging from 22 to 32 m/s. Pull-out force, displacement, and energy absorption over time are quantified for commercially available Kevlar® KM2+ woven fabric with pull lengths L = 5 mm and 27 mm in both quasi-static (QS) and dynamic loading. Dynamic peak pull-out force increases over QS by as much as 290–460% and 170–220% for L = 5 mm and L = 27 mm, respectively, and overall energy absorption increases by 140–420% and 40–140% for L = 5 mm and L = 27 mm, respectively. In QS loading, stick–slip interactions result in forces dropping by approximately 70–80% between oscillating peaks for L = 5 mm and 30–40% between peaks for L = 27 mm, and displacements between peaks are typically 1.5 mm, corresponding to two yarn crossings or two spans. However, at dynamic rates, stick–slip stage force drops by 100% after the first and all subsequent peaks due to the dynamic release of the pull yarn from the fabric, resulting in the pull yarn passing through multiple yarn crossings. To better understand the underlying mechanisms for these observations, finite element simulations of the experiments are performed. Strain rate dependent longitudinal shear stiffness of the yarn is found to significantly affect the dynamic yarn pull-out response and explain the enhanced dynamic yarn pull-out peak force and energy absorption. While the shorter yarn of L = 5 mm experiences a more uniform loading along the entire length, non-uniform loading via progressive uncrimping is observed in the longer yarn of L = 27 mm along the length.