In the present study, fuel cell performance on the MEA model for PEMFC is investigated numerically. The computational domain includes the gas diffusion layer, the catalyst layer and the membrane. The modeling framework is assuming that the transport process is diffusion controlled and the convection transport is neglected. Both the single phase and two-phase flows are studied. In the practical application of the PEM fuel cell, condensed water would exist at high current density. The effect of liquid water is to reduce the diffusivity of the gas phase species. If the pores in a porous media are occupied by liquid water, the gas phase species can not reach the catalyst layer. In past studies, the effective surface area for reactions is modified using the correction factor (1-s). The predicted results, shows that the predicted limiting current is too high compared with the measurements by Liu et al.[5]. To improve the porosity capability, in the present study, the porosity is modified using the correction factor (1-s^(2/3)). The rationale is that because s is defined as the ratio of the volume of liquid water to the volume of void space. When the liquid water is generated by the chemical reaction, it blocks the void cross-sectional area of the porous diffuser, i.e. s^(2/3). Thus, when s increases, the porosity decreases. The predicted IV curves using the adopted porosity modification compare favorably with the measurements. The model is further applied to simulate non-isothermal condition. As expects, the performances of the non-isothermal cases exceed those under the iso-thermal conditions.