How to use survey data to build digital stratigraphic models, assess the uncertainty of models, and provide reasonable survey optimization schemes is one of the important tasks in the digital analysis of geotechnical engineering. This article offers a digital modeling concept based on surface and borehole data, with digital images as the medium. It uses the standard deviation of Kriging prediction to establish the twin-error map of the digital strata, and evaluates the uncertainty of the stratum boundaries using this twin-error map as an indicator of the accuracy of the digital strata. In addition, based on the proposed model concept, a sequential borehole optimization method based on the column error rate indicator is proposed. Case analysis shows that the sequential borehole optimization method, based on the digital stratigraphic model, significantly improves the refinement of the geological model and more effectively reduces the uncertainty of the digital model compared to conventional distance-based borehole arrangements.