The intensification of landscape fragmentation severely disrupts the directional fluxes of information, energy, and matter within ecosystems, thereby compromising ecological networks (ENs) functionality and process integrity. However, traditional EN connectivity optimization predominantly relies on undirected topologies, neglecting the inherent asymmetry and weight heterogeneity of ecological flows-a methodological oversight that creates a critical disconnect between theoretical optimization strategies and actual ecological dynamics. To address this gap, we developed an integrated framework coupling static structural analysis with dynamic directed-flow optimization and applied it to the Sanchuan River Watershed (SRW), a typical ecologically fragile catchment on the Chinese Loess Plateau (CLP). This framework was designed to minimize ecological resistance costs while precisely identifying key nodes and priority corridors. Our results identified 35 ecological sources, 80 potential corridors, 19 pinch points, and 16 barrier points within the SRW. Comparative analysis of static versus dynamic topologies revealed three distinct categories of critical nodes: key origin-destination hubs, pivotal transit nodes, and nodes with previously underestimated importance. Notably, neglecting these "underestimated" nodes in conservation planning scenarios reduced global network efficiency by over 50%. Furthermore, by employing an improved Dijkstra algorithm, we accurately calculated the least-resistance paths between node pairs, significantly optimizing landscape connectivity. This study demonstrates that the explicit incorporation of flow directionality into ENs analysis yields a more realistic representation of ecological processes, effectively rectifying the limitations of traditional static models. The proposed framework offers a robust, operational paradigm for regional EN planning and landscape connectivity optimization in ecologically fragile landscapes, with broad applicability to similar hilly-gully catchments in the CLP and other regions under intensive anthropogenic disturbance.