The ecological models sustained by the multivariate analysis allow the simultaneous study of several variables in a large number of individuals. These are used to reduce to a tangible dimension an amount of information that otherwise would take months to understand, and detect the most relevant variables that would be conditioning the structure of the data. Aim of this work was determining ecological indicators for the management of promising species for the production of forest biomass in the Ecuadorian Amazon. A total of 75 sampling units each one 20 x 20 m (400 m²) were collected in evergreen premontane from stratified random sampling by transepts, and tree species were measured (from ≥ 10 cm to d_1.30), as well as the structural (dispersion of tree crowns, sociological position, basal area, density, value of ecological importance), and environmental variables (altitude, degree of intervention, pH and organic matter). The species resulting from the forest inventory were ordered by multivariate analysis based on the response to different environmental variables. The results of the canonical correspondence analysis were globally significant according to the Monte Carlos test, which gave a good ordination to the sampling units and environmental variables, indicating a strong gradient. According to the sociological position of the 178 inventoried species, only 56 (31.4%) had a continuous vertical distribution. An uneven distribution was found in the number of individuals in the height and diameter classes. From all species Iriartea deltoidea showed greater ecological importance with significant response in function to the altitudinal gradient, whereas Otoba glycycarpa, Pentagonia amazonica and Wettinia maynensis presented multiple answers, which allowed to identify the appropriate conditions for the management of these species. This will allow establishing multifunctional plantation programs for the production of forest biomass.