The most severe loading condition a foundation undergoes is what happens during an earthquake. Thus, the design of foundations in seismic areas needs special consideration compared to the static case. Some researchers have studied the seismic bearing capacity of shallow strip footings, either theoretically or experimentally. Three main approaches are followed in theoretical studies: pseudo-static, pseudo-dynamic and full dynamic analysis. In pseudo-static approach, horizontal and vertical accelerations are applied to the center of gravity of the structure or at the foundation level, and the problem is reduced to a static case of bearing capacity with inclined eccentric loads. In most of these solutions, the inertia of the soil mass is not included. In a pseudo-dynamic approach, the failure surface developed during dynamic condition is assumed to be similar to the one under static loading and the equation of motion is derived from the dynamic equilibrium conditions. In this context, the effect of the earthquake on supporting soil is included in the equilibrium equation. Also, the distribution of earthquake acceleration is included. The full dynamic approach is based on time-history analysis using numerical methods. The majority of the earlier studies are analytic in the context of pseudo-static approach. These studies use several different solution methods such as limit equilibrium, stress characteristics, limit analysis and variational methods. The first studies performed in limit equilibrium context did not include the inertia force on soil (Triandafilidis 1965). Subsequent studies considered the seismic forces both on the structure and on the supporting soil mass (Sarma and Iossifelis 1990; Richards et al. 1993; Kumar and Kumar 2003; Tiznado and Paillao 2014). In most of these cases, the failure surface is assumed to be a constant log-spiral (Fig. 1), the focus of which is at the edge of the footing, which is not supported by the actual performance of foundations. Merlos and Romo (2006) presented a new method in which time-varying inertia forces are applied directly to the building. This method is capable of estimating vertical displacements and foundation tilting by integrating the angular moment equilibrium differential equation. The position of the failure surface was obtained by a minimization process at every time increment. Results showed that under an earthquake loading the potential failure surface moves upward and its length shortens as the seismic accelerations increase, reducing the bearing capacity. Some researchers used the upper bound theorem of limit state analysis (Richards et al. 1993; Dormieux and Pecker 1995; Paolucci and Pecker 1997; Soubra 1999; Ghazavi and Parsapajouh 2006) to study the problem in a more rigorous way.