In recent years, the shrinkage of terminal lakes in Central Asia has triggered environmental degradation and disrupted the supply of critical ecosystem services (ES). Understanding the tradeoffs and synergies among ES in these basins is essential for promoting sustainable regional development and enhancing ecological resilience. This study quantified the spatiotemporal dynamics of five major ES: soil conservation, sand fixation, water yield, carbon storage, and habitat quality using process-based ecosystem models. A Bayesian Network model was constructed to disentangle the effects of natural and anthropogenic influences on ES and identify their dominant drivers. To further characterise ES interactions, we developed a directional tradeoff/synergy strength index that captures both the intensity and directionality of inter-ES relationships. The results revealed spatially heterogeneous tradeoff patterns, with certain ES pairs exhibiting strong conflicts despite sharing similar drivers. Scenario-based optimisation highlighted regional differences in ES priorities, for instance, balancing carbon storage and water yield in the Amu Darya Basin, and managing the interplay of sand fixation, water yield, and soil conservation in the Ili River Basin. These findings provide a decision-support basis for targeted ecosystem management in arid terminal lake regions.