Large-scale modern exhibition venues are more sensitive to uneven foundation settlements, where the spatial distribution of the compressive modulus of the bearing layer is essential in controlling foundation deformations. Conventional engineering survey boreholes provide only a small number of precise compressive modulus geotechnical test values, whereas in-situ testing can provide numerous random cone penetration values. To integrate the data of indoor and in-situ tests, a Bayesian spatial interpolation method of compression modulus is proposed in this study. Our research was conducted as follows. Based on the data accuracy of geotechnical engineering investigation, test data were divided into hard and soft data. A spatial random function was then used to describe the spatial variability of the compression modulus. Next, maximum entropy theory was applied to analyze the uncertainty of the soft data. Based on Bayesian theory, a random field interpolation method was then established to estimate the posterior distribution of the compression modulus of unknown points. Finally, to verify the effectiveness of the proposed method, a Bayesian spatial interpolation method was applied to the spatial variability analysis of the compressive modulus of silty clay in the shallow bearing layer ②$_1$ of Shanghai National Convention and Exhibition Center. Compared with the ordinary Kriging interpolation method, the proposed Bayesian method can integrate multi-source survey data for spatial interpolation with greater accuracy.