Biodiversity is essential for human survival and economic well being and for the ecosystem function and stability. In this study, two classification methods were used simultaneously for determining ecosystem units. 111 plant species in 58 plots were appropriated to four unites: unit 1 (Pyrus syriaca), Unit 2 (Quercus infectoria group), Unit 3 (Q. brantii) and unit 4 (Rosa caniana). Six diversity and evenness indices consist of: Shannon's (H/), Simpson’s reciprocal, Hill's N1, Pielou's J/, Alatalo's F and Molinari's G were calculated. One way ANOVA and Duncan’s multiple comparison tests were used to study differences among ecosystem units. Q. infectoria and R. caniana units had more species diversity than others. Slope, aspect, elevation, trees and herbal percent coverage were used in analyses. Discriminat analysis result to construct discriminant function based on slope percent and Molinari’s index.