The effective stress coefficient α determines the effective stress, which dominating the permeability of rocks. However, the documented value of α for rocks shows a high scatter (0.3-5.5), based on the laboratory measurement. The well know Clay Shell Model (CSM) successfully explain why the α of the clayey sandstone can well above 1 theoretically. However, CSM cannot account for the stress dependency of α observed experimentally. In this study, a modification of CSM was proposed. This proposed Discretized Clay Shell Model (DCSM) discretizes multi-layers clay domain to account for the stress dependent elastic modulus of clay and calculates the pore radius of DCSM model under different confining stress σc and pore pressure Pp. Iso-pore radius curves under different σc and Pp was used to determine the α. The parametric study illustrates the superior features of the proposed DCSM to the traditional CSM. Critical findings include: (1) The predicted α form a concaving upward surface in the pore pressure-confining stress space using DCSM. (2) The influence of Pp on α will be stronger than the influence of σc especially under low Pp. (3) The predicted α is not necessary positively or negatively correlated to σc under constant Pp. (4) The predicted α for soft, high stress dependent deformability of clay coating on the pores of sandstones could be far higher than 1. Two synthetic cases (laboratory and in-situ scale) illustrate the importance of stress dependent (σc and Pp) effective stress coefficient α for determining the effective stress and permeability.