This paper presents an application of the second-order reliability method (SORM) to analyze shallow landslides considered as unsaturated infinite soil slopes. The accuracy of modeling the unsaturated infinite slope reliability depends on the method of evaluating the reliability index and consideration of the soil-water characteristic curve (SWCC). The variability associated with shear strength parameters and fitting parameters of the SWCC of soil is given due consideration for the reliability assessment of unsaturated soil slopes. The formulation for the unsaturated infinite soil slope using the first-order reliability method (FORM) and SORM is presented. The reliability assessment of the unsaturated soil slope based on the Monte Carlo simulations (MCS) method is used as a reference to assess the accuracy of the FORM and SORM. The results demonstrated that the percentage difference between the probability of failures evaluated using SORM and MCS is much smaller than the probability of failures computed using FORM and MCS. The comparative study shows that SORM provides more efficient and accurate reliability indices due to consideration of nonlinear limit state function associated with unsaturated soil slopes. Furthermore, six unsaturated shear strength (USS) models are considered from the literature to evaluate the slope reliability, and the most conservative model is proposed based on the results obtained. The results further revealed that the mean and standard deviations of fitting parameters of the SWCC have considerable influence on the reliability of unsaturated slopes. Finally, design charts for shallow infinite unsaturated soil slopes are proposed, because these are handy to researchers and practicing engineers to evaluate the risk associated with shallow landslides. © 2021 American Society of Civil Engineers.