Understanding how ecological resilience (ER) responds to land degradation (LD) is critical for achieving land degradation neutrality (LDN), particularly in drylands where ecosystems are highly sensitive to hydro-climatic stress and prone to nonlinear shifts. However, empirical identification of resilience thresholds and their implications for management remain limited. Taking the Northern Slope Economic Belt of the Tianshan Mountains-a representative arid socio-ecological system-as a case study, this research develops an integrated framework that links ER dynamics, LDN indicators, and nonlinear threshold detection using restricted cubic spline regression. Results reveal pronounced spatial-temporal heterogeneity. Terrestrial resilience remained low but showed modest recovery from 2000 to 2020, whereas aquatic resilience increased steadily. Overall, the region approached LDN, with net restoration areas expanding from 1.82 & times; 105 km2 to 3.79 & times; 105 km2. Despite this progress, substantial soil organic carbon losses persisted in mountain-oasis transition zones. Nonlinear analysis identified dynamic and multi-stage threshold responses between ER and LD. During 2000-2010, the ER-LD relationship showed multiple inflection points (-2.13%, 56.25%, and 88.06%), indicating alternating phases of sensitivity and buffering. During 2010-2020, the relationship shifted to a clearer inverted U-shaped pattern, with thresholds at 2.51% and 91.96%, suggesting increased system sensitivity and earlier onset of resilience decline. These results highlight the presence of shifting thresholds. Based on these thresholds, two management zones were delineated-Hotspots (1.98 & times; 105 km2), where severe degradation coincided with declining resilience, and Brightspots (0.69 & times; 105 km2), where resilience improved despite environmental pressure. By revealing the evolving, threshold-dependent relationship between ER and LD, this study provides a transferable framework for early intervention, threshold-based management, and resilience-oriented land governance. The findings offer practical insights to support the implementation of LDN and sustainable land management under SDG 15.3 in dryland regions.