The tropical semiarid presents soils with certain degrees of degradation, where the vegetation cannot recover its floristic potential naturally. In addition, the source material of predominant soils in anthropogenically disturbed regions may be determinants of microbial diversity and activity. Therefore, we evaluated the population diversity of rhizobia and arbuscular mycorrhizal fungi (AMF) in soils formed by acidic (Planossolo) and basic (Luvissolo) rock in an environment with and without the predominant vegetation of the Caatinga biome in the state of Pernambuco, Brazil. Chemical characterizations, soil density (DS) and granulometry were performed for each pedogenetic horizon of the studied soils. In parallel, the rhizobia population was determined for cowpea (Vigna unguiculata L. (Walp)), so that after 32 days of inoculation, the plants were collected and the number of rhizobia in each evaluated soil was estimated. Fifteen nodules per horizon were used for isolation and phenotypic characterization, then DNA profiles were obtained in box-PCR analyzes for 132 bacterial isolates, of which 30 strains were selected as representatives and used for genomic DNA sequencing and determination of genetic diversity. Fungal spores were extracted and AMF colonization was determined, using the jurema-preta species (Mimosa tenuiflora) as a host plant, thus obtaining the mycorrhization rate. Soil microbial biomass phosphorus was estimated, in addition to the quantification of glomalin in easily extractable (GFE) and total (GT) protein fractions. Luvissol had higher content of the elements K⁺, Na⁺, Ca²⁺, Mg²⁺ and clay in relation to Planosol, however, the presence of vegetation cover had no influence on the levels of these chemical elements. The rhizobian population was negatively affected by the contents of Mg, Mn, Fe, pH, Ca²⁺, Mg²⁺ and clay, while the number of rhizobian isolates was affected by the presence of Ca, Zr, Ca²⁺, K⁺ and clay. Rhizobian diversity was affected by K₂O, K⁺ and Ca²⁺. All soils and their pedogenetic horizons showed bacterial isolates capable of solubilizing calcium phosphate, however, the rhizobian population was higher in Planosol in relation to Luvissol and the presence of vegetation favored a greater bacterial diversity, which may be more apt to functional redundancies, even in soils considered less fertile. Bacterial isolates belong to the genus Paenibacillus, Bacillus, Aneurinibacillus, Priestia and Methylobacterium, with a predominance of bacteria belonging to the phylum Fimicutes, which are capable of symbiosis, making P and K available to plants, or being associated with plant roots, when still present in non-degraded soils. GFE and GT were significantly higher in Luvissolo, whereas phosphorus in the microbial biomass was significantly higher in Planossolo, without the influence of vegetation cover. The number of fungal spores showed a negative correlation for Ca, Mg²⁺, Clay, Mg, Ti, Fe, Na⁺ and Ca²⁺, whereas the colonization rate showed a positive correlation for P and K⁺. Although more fertile than Planosols, Luvissols had a greater amount of GFE and GT proteins than Planosols, but there was no development of AMF spores, regardless of the presence or absence of vegetation. In general, both soils studied still have bacterial and fungal populations capable of fixing nitrogen and making P available, respectively, which can be the target of ecological succession studies in potentially degraded environments, even in the deepest horizons.