Terrestrial plants diversity is greatest in the tropics, therefore, understanding plant diversity patterns and processes in the tropics is critical to global biodiversity conservation efforts. However, the tropical region has and continues to experience a drastic change in the last century, where most of its forest cover had been cleared. As forests are cleared, areas that were once contiguous forest have become a landscape of smaller and isolated forest areas (forest fragments), a process known as forest fragmentation. While there is a wealth of knowledge on ecological consequences of forest fragmentation on plant compositional and diversity patterns, most of these studies are centered in South and North America. In contrast, the effects of fragmentation in tropical Asia are much less researched. This is of concern as the Asian tropical region has experienced the highest proportional loss of forest of all tropical regions. Studies of vegetation dynamics in tropical forests have focused historically on larger and more economically important tree species, leaving patterns of understory herbaceous diversity poorly understood. The distribution and abundance of understory species in tropical forest has been correlated with a variety of historical factors and complex environmental gradients, especially those related to soil condition, water and light availability across the landscape and over time. This study presents plant compositional and diversity patterns from a forest fragmented landscape in Xishuangbanna Prefecture, Yunnan Province, China. I focus on two topics: (1) how well trees represent diversity patterns and responses to fragmentation of other plant lifeforms, and (2) whether diversity of different understory herb functional groups respond differently to water-fertility and water-light resource gradient trade-offs. To achieve these topics, we sampled different plant life-forms from 50 fragment plots in Xishuangbanna, and we used compositional and diversity patterns of 861 identified plant species where 446 were trees, 176 were lianas and 239 species were understory herbs and ferns. I addressed the first topic by conducting separate analysis for trees and non-trees (lianas, herbs, and ferns). I compared community types generated by clustering analysis for the nontree life-forms with those generated for trees. Cluster analysis generated three distinct vegetation type, namely: lowland forest, montane forest and limestone forest. I found that plot-level tree data can represent vegetation community types of non-tree life-forms, but is poorly representative of the richness and density of other life-forms, and poorly represents forest fragmentation responses for the entire plant community. This means that other plant life-forms may not be well represented when drawing conservation plans and management strategies that address the needs of trees only, suggesting the importance of considering broader functional groups when assessing vegetation responses to altered environments. Fragmentation responses of different plant life-forms differed between community types, which suggests the importance of recognising distinct plant communities within a landscape to identify where fragments should be prioritised for conservation planning. I addressed the second topic by looking at different understory herbaceous life-forms, namely ferns, eudicots and monocots, and checked whether their compositional and diversity patterns they are driven by different resource trade-offs, specifically water-fertility and water-light predictors taken from the study area. Compositional patterns in eudicots were more greatly linked to water-fertility than to water-light trade-offs while monocots and ferns utilised both trade-offs equally. Diversity patterns of eudicots, monocots and ferns were consistently more strongly linked to water-light than to water-fertility trade-offs. The results provide evidence of habitat filtering on composition and diversity patterns of co-occurring ferns, dicots and monocots, strengthening the role of environmental factors in shaping species assemblage patterns. This thesis provides substantial evidence of forest fragmentation and the importance of environmental parameters in driving plant compositional and diversity patterns in Xishuangbanna, Southwest China.