Vertical drains have been extensively used as a liquefaction countermeasure and seismic case histories have demonstrated the effectiveness of these technique. However, it has often been observed that excessive settlement and deformation of the ground occurred even though excess pore water pressure was expected to dissipate and be kept low by the technique. Uncertainties and questions remain regarding the use of these technique as liquefaction remediation. In this study, a series of centrifuge tests were conducted to promote better understanding of the performance of sand remediated with gravel drains. The effects of drain diam-eter/spacing ratio, and permeability and depth of liquefiable layer were systematically investigated. Excess pore pressures observed in the tests are also compared with those predicted with the Seed and Booker’s design procedure. It is confirmed that the use of a constant mv value over the depth in the current design procedure makes the design somewhat unsafe.