<p>Agricultural practices threaten environmental sustainability at both local and transboundary scales, yet how to optimize them for improved sustainability across these multiple scales remains poorly understood. Here we propose an integrated framework combining safe and just operating spaces, spatial optimization, and extended multiregional Input-Output model to optimize crop-livestock systems on the Qinghai-Tibetan Plateau for sustainable internal–external outcomes. We show that 6.5% of the Qinghai-Tibetan Plateau has exceeded three or more environmental boundaries, but 56.1% of these could be alleviated by reorienting crop and livestock distributions. Such optimizations could reduce greenhouse gas emissions (8.7%) and phosphorus surplus (3.0%), increase agricultural output by 9.5%, and save water resources to benefit downstream areas (1.3%–74.5%), while realizing a reduction of 0.14 megatons of CO₂-equivalent, 1.21 kilotons of nitrogen, and 0.22 kilotons of phosphorus emissions outside the Qinghai-Tibetan Plateau. This framework underscores the importance of integrated spatial planning for achieving sustainability at local and transboundary scales.</p>