This study is aimed at understanding anthropogenic impact on the understory plant diversity around the municipal area of Xiamen in Fujian Province, SE China. Four indexes were employed to differentiate the diversity of urban forests adjacent to different land use types. Principle Component Analysis (PCA) and Canonical Correspondence Analusis (CCA) methods were applied in our study. The results showed that human trample, dustheap, digging, accessibility and other factors have great influence on the understory plant diversity. Herb β diversity of urban forest increased along the land use gradient from commercial, transportation, industrial to residential, while shrub β diversity decreased gradually along this environment gradient.