As a remedial approach, vacuum-induced pore-water extraction offers the possibility of contaminant and water removal from the vadose zone, which may be beneficial in reducing the flux of vadose zone contaminants to groundwater. Vadose zone water extraction is being considered at the Hanford Site in Washington State as a means to remove technetium-99 contamination from low permeability sediments with relatively high water contents. A series of intermediate-scale laboratory experiments have been conducted to improve the fundamental understanding and to recognize the limitations of the technique. Column experiments were designed to investigate the relations between imposed suctions, water saturations, and water extraction. Flow cell experiments were conducted to investigate the effects of high-permeability layers and near-well compaction on pore-water extraction efficiency. Results show that water extraction from unsaturated systems can be achieved in low permeability sediments, provided that the initial water saturations are relatively high. The presence of a high-permeability layer decreased the yield, and compaction near the well screen had a limited effect on overall performance. In all experiments, large pressure gradients were observed near the extraction screen. Minimum requirements for water extraction include an imposed suction larger than the initial sediment capillary pressure in combination with a fully saturated seepage-face boundary. A numerical multiphase simulator with a coupled seepage-face boundary condition was used to simulate the experiments. Reasonable matches were obtained between measured and simulated results for both water extraction and capillary pressures, suggesting that numerical simulations may be used as a design tool for field-scale applications of pore-water extraction.