In order to monitor plant diversity, yearly census of plant species is required at given locations. However, there are several methods to record the species. Apart from well-established methods such as Braun-Blanquet scale, plants are surveyed by their density, coverage, and also by their frequency. The present study compared the results of these parameters for the evaluation of diversity, based on observations carried out on Mount Usu, an active volcano (Hokkaido, Japan, 42°32’N, 140°50’E). Herbaceous plants were surveyed at 140 quadrats (1 m*1 m) during 2015-17. The quadrats were chosen by stratified random two-stage sampling, and covered four different forest types in various succession stages: closed and open broadleaf, semi-artificial and artificial forests. Diversity was calculated according to the true diversity index when applicable. The results showed that the species diversity followed similar trend across the methods, but differed in magnitude, which lead to over- or underestimation of diversity. Further comparison also indicated great variation in plant sizes and the dominance of large species in the quadrats. The results help to identify the best parameter or combination of parameters to monitor diversity and to manage frequently disturbed protected areas, such as volcanic landscapes.