Richness, diversity, and evenness indices of species under three level of utilization (reference, key, and critical areas) in different land units ( plain, hill, and mountain) were compared by 10xl belt transect. Form richness indices, the number of all species in the sample (NO) and richness based on rarefaction method had the same behavior at different level of utilization, which were quite fitted to the real situation. But Margalef and Menhinick indices (R1,R2) did not show any sensitivity to the different levels of utilization, and were considered as improper indices. Richness based on long normal models also exaggerated the actual number of species and was not recommended. Form diversity indices, Shannon's index was preferred to the other indices which it was sensitive to the rare species at reference and key areas. So the effects of protection in terms of rare species can be evaluated by this index. Because of occurrence of rare species in reference and key areas and invaders in critical area, evenness index was higher at critical areas, so among evenness indices, Pets index (E1) was close to one, so it was more suitable than the others.