Abstract: Paleocurrent analysis plays an important role in the fields of sedimentary petrology and paleogeographic reconstruction. Traditional methods of field measurement for paleocurrent data are plagued by low work efficiency and blind spots in data acquisition. With the advancement of unmanned aerial vehicle (UAV) technology and geographic information systems, real-scene digital outcrop 3D models have provided a new method for the measurement and analysis of paleocurrents. This research is based on 3D models constructed using UAV oblique photography, which provide high-resolution texture information that allows geologists to accurately identify sedimentary structures which indicate paleocurrent directions. Utilizing the precise spatial information from 3D models, a platform for real-scene 3D models of digital outcrops is built for the visualization of the models. Paleocurrent measurements of commonly observed cross-bedding in the field are quantitatively carried out through a spatial three-point method. Computer correction of paleocurrent data is conducted based on azimuthal equidistant projection and spatial geometric relationship correction algorithms to eliminate the effects of stratigraphic structural changes. The processed paleocurrent data are subjected to statistical analysis and visual presentation. Interactive methods are used to draw regional rose diagrams, providing an intuitive display of the distribution and dominant directions of paleocurrent data within the region. Furthermore, this paper uses the Guanjiaya outcrops in the Ordos Basin as examples to implement sedimentary paleocurrent analysis based on real-scene digital outcrop 3D models. Through the aforementioned methods, the analysis of paleocurrents effectively replaces paleocurrent data field measurement and reduces the fieldwork burden. Additionally, It provides data support for interpreting source directions and inferring ancient slopes.