Automatic waterway thematic information from high-resolution images helps GIS database update and facilitates waterway management and maintenance. In this study, waterway manmade objects are extracted by object-based image analysis (OBIA), which can utilize abundant features and flexible rule sets based on object primitive. However, routine OBIAs often follow a region-based framework, i.e., the technical route of ??segment and then classify?, which limits their performance on manmade object extraction. In previous studies, we established a new OBIA technology model called the region-line primitive association framework (RLPAF) that involves both regions and straight lines. In RLPAF, straight lines are detected aside from regions. Straight line and region-line association features are extracted. In the object classification stage, regions and lines are collaboratively utilized for comprehensive image analysis which makes OBIA flexible and improve its performance. We developed several RLPAF-based object extraction algorithms for HSR images and validated that RLPAF performed better in man-made object extraction compared with conventional region-based OBIA technology. In this study, RLPAF is applied for thematic information extraction along the Yangtze River Waterway from HSR images. The technical route includes image segmentation, object-based land/water classification, shoreline buffering, region-line association feature extraction, object extraction based on RLPAF rules, and object refinement involving straight line primitives. The experimental results show that the proposed techniques and methods can extract several kinds of waterway object, e.g., wharves and bridges, from HSR images with better accuracy than routine OBIAs.