The aim of this research was to evaluate variability among BRS 8381 soybean yield components, under different sowing densities and two amounts of limestone applications on a savanna in Roraima. The seeds were sown to provide five plant populations (10, 14, 18, 22 and 26 linear m -1 ), in plots with the application of 1.5 and 4.5 t ha -1 of limestone. Experimental plots consisted of four 5-meter long do 5 rows spaced 0.50 cm from each other. The experimental design was of randomized blocks in a factorial scheme with four replications. To evaluate the proposed variables, plants placed within the measurable area of each plot were collected and the following parameters were assessed: plant height, height of insertion of the first pod, stem diameter, number of nodes on the main stem, number of stems per plant, total number of pods per plant, number of grains per pod. Of the total number of plants harvested, 20 were randomly selected and evaluated individually. In addition to the aforementioned parameters, number of pods on the four upper nodes of the plant, dry mass of the plant and dry mass of grain were also evaluated. The apparent harvest index, 100-grain mass and yield estimate were also obtained. The BRS 8381 cultivar grown with 10 plants per linear meter of row presented greater plant dry mass, dry mass of grain, dry mass of hull and number of pods and twigs in both environments. The maximum grain yield was obtained with a plant distribution of 22 plants m -1 of row. The cultivar BRS 8381 is recommended with liming at 4.5 t ha -1 of dolomitic limestone under the edaphoclimatic conditions of the present study, due to its high phenotypic plasticity and higher performance in the production components.