The magnitude and pace of anthropogenic climate change are increasing the likelihood of abrupt changes in terrestrial ecosystems worldwide. The Last Deglacial Transition (LDT) witnessed global and regional climate change and caused vegetation change globally. The recognition and attribution of deglacial abrupt vegetation conversion would certainly enhance our understanding on the survivability and resilience of ecosystems to climate change. Here, we conducted statistical analyses for a pollen record from Yilong Lake, southwestern China, focusing on the period of the LDT. Biome changes since the Last Glacial Maximum were reconstructed numerically by using a standard approach of pollen-plant functional type-biome assignment (biomization). Principal component analysis and structure change test were applied on pollen composition and diversity data. Modern vegetation and climate relationship was analyzed to discuss climatic conditions for the vegetation conversion in Yilong region during the LDT. Our results revealed a vegetation conversion occurring around 14.3 cal kyr BP, characterized by abrupt shifts in pollen compositions and biomes and also reduction in diversity indices. We suggested that the cumulative pressure of climate warming during the LDT was probably the critical climatic variable triggering the vegetation conversion in Yilong region, while precipitation played an important role in regulating the proportions or coverages of different vegetation types after the development of evergreen broadleaved forest. By the end of tthis century, climate under intermediate to very high greenhouse gas emission scenarios were projected to reach the thresholds for vegetation transitions from evergreen broadleaved forest to (monsoon) rainforest, or to Savanna-like vegetation. However, it is still a challenge to forecast accurately when or on what spatial scale vegetation change will happen because vegetation responses to climate change are complicated and the topography and habitats on Yunnan Plateau are extremely diverse.