Abstract: Plant functional types (PFTs) bridge the gap between plant physiology and ecosystem processes, providing a powerful tool for studies on global change, vegetation dynamics and vegetation-atmosphere processes. In this paper, a two-stepped hierarchical PFT classification system was proposed based on six key plant attributes. Three of six key attributes were related to plant canopy structure, including permanence of aboveground live biomass (woody vs. herbaceous), leaf longevity (evergreen vs. deciduous) and leaf structure (broad-leaved vs. needle-leaved). The other three attributes were related to plant physiological characteristics, including photosynthetic pathway (i.e. C3 vs. C4 grasses), drought resistance and temperature tolerance (e.g. warm evergreen, cool conifer). In the first step of the analysis, five basic plant types were derived based on canopy attributes. In the second step, these five basic plant types were sub-divided into 29 types according to their photosynthetic pathways, drought resistance and temperature tolerance. We tested this PFT classification system by simulating the distribution of vegetation in China. A sub-set of 18 plant types was selected from the 29 PFTs that were characteristic of the vegetation in China, especially in relation to the water balance and energy budget as affected by the monsoon climate and the Tibetan Plateau. This set of PFTs contained 7 trees, 6 shrubs and 5 grasses. Two types of bare ground were added for the simulation. Six climatic variables were selected to be used as the distributional constraint of each PFT, which included absolute minimum temperature (Tmin), mean temperature of the warmest month (Tw), growing-day degrees (GDDs), annual range of monthly mean temperature (DTY), moisture index (annual precipitation-annual potential evapo-transpiration, MI), annual precipitation (P). The values of these climatic variables were determined based on the geographical distribution of each plant functional type. Using BIOME1, the distribution of the PFTs across China was then modeled. The simulated map of China's vegetation was in close agreement with the actual vegetation map, indicating that this set of PFTs was capable of simulating the geographical distribution of vegetation in China. This study provides a basis for future studies on climate-vegetation interactions and for developing regional dynamic vegetation models and regional climate models for China.