Abstract: To obtain more valuable site-specific test data with the minimal site exploration effort, it is of great significance to design an optimal site exploration program before geotechnical site investigation. This paper aims to propose an optimization design approach for layout scheme of slope boreholes based on the Bayesian updating and value of information analysis. The BUS (Bayesian updating with structural method) with subset simulation is employed to update the posterior statistical properties of spatially varying soil properties and to estimate the posterior probability of slope failure. Then, the value of information analysis is used to determine the most optimal borehole locations and separation distances between two boreholes. Additionally, to accurately characterize the prior information of soil parameters, a non-stationary random field model is developed wherein the depth-dependent nature of soil parameters is addressed. Finally, a clay slope example under undrained conditions is investigated to demonstrate the effectiveness of the proposed approach. The results indicate that the proposed approach can effectively determine the most optimal borehole locations and separation distances between two boreholes prior to performing site investigation tests based on the prior information of soil parameters. To design a separation distance between two boreholes as small as possible during geotechnical site investigation is not necessarily the best choice to understand the performance of the slope.