Assessing the supply and demand balance (SDB) of ecosystem services (ESs) in arid zone is crucial for sustainable development of ecologically fragile regions and land-space optimization. The present research takes Xinjiang as a case study and, based on three-phase data from 2000 to 2020, evaluates the SDB of five key ESs-through the application of the ESs supply and demand ratio (ESDR). The ordered weighted averaging (OWA) algorithm is then employed to simulate multi-scenario decision-making, identify priority protected areas (PPA), define land use (LUCC) thresholds required to achieve SDB of ESs, and develop a zoning control scheme. Results show that, except for CS, the overall supply of the other four ESs exceeds demand but their ecological surplus declines over time. The S4 scenario (risk = 0.3) is identified as the optimal scheme, under which 322,000 km2 of PPA are designated, mainly in the Kunlun, Tianshan and Altai Mountains. Cultivate land, construction and forest land are identified as key LUCC types affecting ESs balance, and LUCC thresholds are determined for different services. At the 5 km grid scale, three types of control units are delineated: GP guarantee zone, HQ-CS-SR comprehensive guarantee zone, and ESs improvement zone. This study develops a "threshold-partition" linkage LUCC regulation path and a management framework for PPA that supports refined land space management and ecological protection and restoration in arid zone.