This study focused on the impact of land-use changes and agricultural management of soybean in Amazon forest soils on the abundance and composition of the acidobacterial community. Quantitative real-time PCR (q-PCR) assays and pyrosequencing of 16S rRNA gene were applied to study the acidobacterial community in bulk soil samples from soybean croplands and adjacent native forests, and mesocosm soil samples from soybean rhizosphere. Based on qPCR measurements, Acidobacteria accounted for 23% in forest soils, 18% in cropland soils, and 14% in soybean rhizosphere of the total bacterial signals. From the 16S rRNA gene sequences of Bacteria domain, the phylum Acidobacteria represented 28% of the sequences from forest soils, 16% from cropland soils, and 17% from soybean rhizosphere. Acidobacteria subgroups 1–8, 10, 11, 13, 17, 18, 22, and 25 were detected with subgroup 1 as dominant among them. Subgroups 4, 6, and 7 were significantly higher in cropland soils than in forest soils, which subgroups responded to decrease in soil aluminum. Subgroups 6 and 7 responded to high content of soil Ca, Mg, Mn, and B. These results showed a differential response of the Acidobacteria subgroups to abiotic soil factors, and open the possibilities to explore acidobacterial subgroups as early-warning bioindicators of agricultural soil management effects in the Amazon area.