The tall grass vegetation in the Andean grassland ecosystems covers the largest area compared to other types of vegetation such as Puna grass, wetland and others. The grasslands are frequently set on fire by livestock farmer, seriously affecting the ecosystem. One way to mitigate this problem is to use these species as a source of plant fibre, which can be economically useful to the interests of the livestock family without affecting the ecosystem. To advance in this approach, it is necessary to know the functional characteristics of the plants; therefore, we evaluated the aerial primary productivity, plant density per m2, basal cover, aerial cover and leaf height, whose data were analysed using the generalised linear mixed model and the correlation between these variables with the physical-chemical characteristics of the soil, by means of principal component analysis and canonical correlation, in seven species of grassland and seven control plots, located between 3860 and 4333 metres above sea level. The results showed significant differences for p=0.001 between species, and between plots, and a canonical correlation grouped in two clusters that showed the differentiated importance of soil elements with the phytomass produced.