Antimony (Sb) and copper (Cu) are toxic heavy metals that are associated with a wide variety of minerals. Sb(III)-oxidizing bacteria that convert the toxic Sb(III) to the less toxic Sb(V) are potentially useful for environmental Sb bioremediation. A total of 125 culturable Sb(III)/Cu(II)-resistant bacteria from 11 different types of mining soils were isolated. Four strains identified as Arthrobacter, Acinetobacter and Janibacter exhibited notably high minimum inhibitory concentrations (MICs) for Sb(III) (>10 mM),making them the most highly Sb(III)-resistant bacteria to date. Thirty-six strains were able to oxidize Sb(III), including Pseudomonas-, Comamonas-, Acinetobacter-, Sphingopyxis-, Paracoccus- Aminobacter-, Arthrobacter-, Bacillus-, Janibacter- and Variovorax-like isolates. Canonical correspondence analysis (CCA) revealed that the soil concentrations of Sb and Cu were the most obvious environmental factors affecting the culturable bacterial population structures. Stepwise linear regression was used to create two predictive models for the correlation between soil characteristics and the bacterial Sb(III) or Cu(II) resistance. The concentrations of Sb and Cu in the soil was the significant factors affecting the bacterial Sb(III) resistance, whereas the concentrations of S and P in the soil greatly affected the bacterial Cu(II) resistance. The two stepwise linear regression models that we derived are as follows: MIC(Sb(III))=606.605+0.14533 x C(Sb)+0.4128 x C(Cu) and MIC((Cu)(II))=58.3844+0.02119 x C(S)+0.00199 x CP [where the MIC(Sb(III)) and MIC(Cu(II)) represent the average bacterial MIC for the metal of each soil (μM), and the C(Sb), C(Cu), C(S) and C(P) represent concentrations for Sb, Cu, S and P (mg/kg) in soil, respectively, p<0.01]. The stepwise linear regression models we developed suggest that metals as well as other soil physicochemical parameters can contribute to bacterial resistance to metals.