The climate-ecological mechanisms driving vegetation dynamics in the ecologically fragile ecosystems of Central Asia remain insufficiently understood, particularly under accelerating climate change. To address this, we developed an integrated analytical framework (TMMS) combining trend tests, multiple regression, and structural equation modeling to systematically assess Leaf Area Index (LAI) dynamics and their driving mechanisms from 2001 to 2023. Results reveal distinct longitudinal and latitudinal gradients, alongside a significant vertical 'increase-then-decrease' pattern driven by topographically controlled hydrothermal conditions. Temporally, it underwent 'increase-decline-recovery' fluctuations. Vegetation degradation was primarily concentrated in grasslands during the summer. Soil moisture was the key factor governing these spatiotemporal variations. Further analysis uncovered that LAI dynamics are regulated by a multi-pathway mechanism of 'climatic forcing-soil moisture regulation-carbon-water coupling-biomass accumulation'. Under drought stress, vegetation exhibits a 'survival-first' adaptive strategy, where carbon use efficiency negatively correlates with biomass, thus prioritizing metabolic maintenance over growth. Future trend predictions indicate that typical water source areas and oasis-edge zones possess significant ecological restoration potential, while the degradation risk in water-depleted areas remains severe. The proposed TMMS framework provides a robust tool for understanding complex vegetation responses and offers scientific guidance for sustainable management in arid regions.