Deformation of soil layers during earthquake shaking is a major cause of damage to buildings and geotechnical structures. Over the past 40 years, several design methods have been proposed to quantify and estimate the settlement of sand layers due to seismic loads. Most of these methods have been focused on sands in dry or water-saturated conditions, leaving a gap in the basic understanding of the mechanisms of seismically induced settlements of partially-saturated sands. Design approaches for partially-saturated soil layers deserve further development because these soil layers may have different deformation mechanisms during earthquake loading than water-saturated or dry soil layers. In this research an effective stress-based empirical methodology is proposed to predict the earthquake induced settlement of a free field partially-saturated sand layer. This approach uses a complex coupling between stress state, seismic compression, and pore water pressure generation during earthquake loading. Accordingly, an experimental study on partially saturated soils was used in tandem with this design model to quantify input relationships for the different variables during earthquake loading and also to verify the induced settlement estimated using the proposed empirical approach. A new centrifuge physical modeling technique was developed that involves a soil specimen within a laminar container mounted atop a hydraulic servo-controlled shake table in a geotechnical centrifuge. Steady state infiltration of water was used to control the stress state in the partially-saturated sand layer during centrifugation. Specifically, infiltration was found to lead to a relatively uniform degree of saturation with depth in the sand layer, simplifying interpretation of the deformation response. Sand layers with a wide range of degrees of saturation were evaluated using this approach to assess the degree of saturation on their deformation response during cyclic loading. A nonlinear trend was observed in the variation of the surface settlement with degree of saturation, with a minimum value obtained for sand with a degree of saturation of about 0.4. This trend is consistent with the relationship between small-strain shear modulus and degree of saturation measured using bender elements and resonant column testing techniques and also with the predicted trend using the developed empirical methodology.