To enhance ecological resilience in environmentally fragile restoration zones, this study selects Yanchi County in Ningxia, a representative eco-restoration area in northern China, as a case study. A three-tier analytical framework - comprising ecological function identification, process simulation, and resilience-oriented regulation - is developed. By integrating Multi-scale Patching and Spatial Analysis (MSPA), dynamic resistance surface modeling, and circuit theory, ecological sources and corridors are accurately identified. A dual-dimensional resilience assessment system based on the "source-corridor" structure is constructed to examine the spatiotemporal evolution of the ecological network from 2000 to 2020. The results reveal that: (1) Ecological service functions increased by 22.45% from 2000 to 2020, although growth slowed after 2010. Meanwhile, the Ecological Risk Index (ERI) rose by 12.5%, with high-risk zones expanding from 14% to 16.3% and shifting northward by 12.6 km, indicating a mismatch between functional gains and risk control; (2) The number of ecological sources decreased by 17% between 2000 and 2010 but rebounded by 8.8% from 2010 to 2020. However, fragmentation intensified, as evidenced by a 21.7% decrease in the largest patch index (LPI), leading to disrupted ecological stability and reduced resilience of source areas;(3) An ecological security pattern - "one core, two zones, two belts, and multiple corridors" - is proposed to enhance network connectivity and mitigate regional ecological threats, particularly wind erosion and soil degradation. This study provides a scientific and technical framework for optimizing ecological security patterns from a resilience perspective in arid and semi-arid regions, offering empirical support for landscape-scale ecological restoration and sustainable management.