Abstract: The inherent variability of soil properties has currently received considerable attention in slope reliability analysis. However, the geological uncertainty is often ignored in analysis. This paper proposes a method for the slope reliability analysis considering geological uncertainty and spatial variability of soil parameters. The coupled Markov chain model is adopted to simulate the geological uncertainty. The midpoint method based on the Cholesky decomposition technique is used to discretize the random fields of different soil type parameters. The safety factor is calculated using the finite element-strength reduction method. The slope reliability analysis is conducted by Mote Carlo simulation. The procedure for this method is presented. A slope reliability problem is analyzed using the borehole data in Perth, Australia to illustrate the importance of considering geological uncertainty and spatial variability of soil parameters in slope reliability analysis and investigate the effect of layout scheme of boreholes on evaluating slope reliability. The results indicate that the proposed method can effectively reflect the effect of two types of soil heterogeneity on slope reliability. When there is less number of boreholes, the difference between the simulated and the real soil type distribution of slope is large, and ignoring the geological uncertainty will result in inaccurate estimation for slope reliability. The layout scheme of boreholes has a significant effect on safety factor and failure probability of slope. Boreholes should be placed as many as possible in the critical influence zone of the slope. The relationships between the statistics of safety factor and failure probability of slope and the number of boreholes are not monotonic, but the statistics of safety factor and failure probability of the slope converge to the accurate value with the increasing borehole number.