To reveal the effects of plantations on soil microbial environment,the composition and diversity of soil fungi and bacterial communities in five restoration models (Robinia pseudoacacia, Populus hopeiensis, Pinus tabuliformis, Picea crassifolia, natural restoration) in the mountainous area of southern Ningxia were compared by using high-throughput sequencing technology. The correlation between soil physical-chemical properties and dominant microbial groups was analyzed. The results showed that: 1) Dominant fungi in different restoration models were Ascomycota, Basidiomycota, Mortierellomycota, and unclassified fungi, which accounted for 90% of total fungal community. The dominant soil bacteria were Actinobacteria, Proteobacteria, Acidobacteriota, Chloroflexi, and other bacteria, accounting for more than 80% of total bacterial community. 2) The diversity of soil fungi in P. tabuliformis forest was the highest, with Shannon index, and Simpson index being 3.72±0.37 and 0.07±0.04, respectively. The richness of fungi in naturally restored forest land was the highest, with Ace and Chao1 index of 708.19±137.25 and 706.26±125.34, respectively. The bacterial diversity and richness of species in P. tabuliformis forest land was the highest. The Shannon, Simpson, Ace and Chao1 indices were 6.57±0.04, 0.004±0.00, 3439.81±41.67, 3463.14±32.16, respectively. 3) The fungus with significant difference among restoration models were Solicoccozyma, Cladosporium, and Alternaria. Bacteria from Norank_f_67-14, Rubrobacter_f_Rubrobacteraceae, Sphingomonas_f_Sphingomonadaceae had significant difference among restoration models. 4) The RDA ordination of the dominant microbial flora and soil physical-chemical properties showed that soil bulk density (BD), carbon to nitrogen ratio (C/N), and pH were the major factors affecting the dominant fungal flora. BD, nitrogen to phosphorus ratio (N/P), total phosphorus (TP), and total carbon (TC) were the main factors affecting the dominant bacterial flora. In general, the difference of composition and diversity in the fungal community of different restoration models was higher than that of the bacterial community, indicating that the fungal communities were more sensitive to the changes of tree species and soil environment than bacterial communities. Our results could provide the theoretical foundation for vegetation restoration measures and the maintenance of ecosystem function stability in southern Ningxia.