Abstract: The Mu Us Desert, an ecologically vulnerable region in northern China, faces severe challenges, including intensified landscape fragmentation, shrinking ecological land, and threats to biodiversity. Focusing on landscape connectivity, this study investigated the landscape ecological network of the desert based on the morphological spatial pattern analysis (MSPA) and circuit theory, aiming to alleviate ecological conflicts and promote regional sustainable development. Specifically, by analyzing the grassland landscape structure through MSPA and combining the probability of connectivity (PC) index, ecological source areas were identified. Then, a comprehensive resistance surface was constructed using resistance factors across three dimensions: nature, society, and landscape pattern. Subsequently, ecological corridors and ecological pinch points were identified based on the circuit theory, and corridor importance was classified in combination with a gravity model. Finally, the current landscape ecological network of the Mu Us Desert was identified and optimized under three scenarios. The results indicate that the current landscape ecological network of the Mu Us Desert comprises 22 ecological source areas, 38 ecological corridors, 27 ecological pinch points, and 52 ecological breaking points, exhibiting spatial heterogeneity characterized by high density in the western and southeastern part and low density in the northeastern part. Network structure indices (α=0.25, β=1.41, γ=0.51) reveal poor closed loops, limited connectivity, and low resistance to disturbances within the Mu Us Desert. Among the three scenarios, scenario 1 (spatial pattern optimization) enhances the east-west spatial balance, effectively improving the landscape ecological network (Δα=0.06, Δβ=0.12, Δγ=0.04). Scenario 2 (buffer zones set around ecological source areas) increases landscape connectivity, demonstrating significant improvement in the landscape ecological network (Δα=0.06, Δβ=-0.08, Δγ=0.06) at lower costs. In contrast, scenario 3 comprehensively enhances the ecological network balance and connectivity (Δα= 0.22, Δβ=0.26, Δγ=0.16), demonstrating the most significant improvement in the landscape ecological network at high costs.