<p>Existing studies primarily construct flood security patterns based on the static "patch-corridor-matrix" research paradigm. This study integrates this paradigm with source-sink theory, proposing a new approach from the "source-flow-sink" landscape perspective. This method combines dynamic stormwater runoff processes with static landscape patterns to create a novel flood security pattern. Using Changsha City as a case study, the research applied the Curve Number Model, iso-volume algorithm, and spatial statistics tools to identify key source-sink landscape patches. Based on the distribution of these patches and existing river system, hydrological analysis tools were used to extract significant flow landscape corridors. Key sub-watersheds were subsequently identified, completing the flood security pattern construction. Finally, the influence of source-sink landscape patterns on flood risk was analyzed using the Spearman coefficient, and control strategies were proposed. The results indicated 29 important source patches with an area of 20.72 km², 49 important sink patches with an area of 110.43 km², and 18 significant flow corridors within the study area. The key sub-watersheds included the Longwanggang, Guitang, Baqu, and downstream reaches of the Liuyang River. Control strategies focused on optimizing surface runoff and improving spatial layout in key source-flow-sink landscapes. Identifying critical source-flow-sink landscapes, based on the flood risk formation mechanism and management requirements, is essential for optimizing watershed landscape patterns and ensuring urban flood security.</p>