Abstract: The migration trajectories of historical literati are a key carrier for studying the sociocultural mobility and regional interaction in ancient China, and their spatiotemporal patterns are deeply linked to individual destinies and the changes of the times. Aiming at the problems of existing large-scale literati migration trajectory research, such as path overlapping leading to visual clutter, insufficient tools for in-depth association analysis, and low interactivity, we collaborated with experts in ancient literature and proposed a large-scale visual analysis system, LiteratiPath, for the migration of literati in the Song Dynasty. First, an Adaptive Edge Path Bundling (AEPB) algorithm was designed to optimize the visual representation and efficiency of large-scale migration trajectories, revealing the macroscopic flow patterns and key paths of group migration data. Then, a four-dimensional data association model of "city-time-literati-poetry" was constructed, integrating elliptical kernel density analysis, migration index calculation, and poetry topic mining techniques to realize the quantitative calculation of migration dynamics and the mining of spatial topic associations. Finally, visual symbols and a multi-view collaborative interaction mechanism were designed to enable progressive exploration of migration patterns. Quantitative analysis was conducted on the Song Dynasty literati migration dataset, and case studies were carried out with 12 literary researchers. The results show that compared with other algorithms, the AEPB algorithm optimizes visual quality, efficiency, and bundling effects, and performs best in terms of distortion, and ambiguity; in user experiments, the proposed method and system provide new analytical perspectives and tool support for historical humanities research.