To investigate the impacts of crude oil pollution on the soil microbial community, and estimate the potential for crude oil degradation by indigenous microbial consortia, we examined the soil microbial community structure, metabolic characteristics and functional diversity of crude oil-contaminated soil collected in the Loess Plateau in northern Shaanxi, using plate counts and the Biolog Eco plate method. The results showed that the responses of soil microbes to crude oil pollution stress varied greatly. The abundance of bacteria and fungi in crude oil-contaminated soil were about one order of magnitude higher than in the uncontaminated soil, while the abundance of actinomycetes was significantly lower in polluted soil than in uncontaminated soil (P < 0.01). The number of bacteria in crude oil-contaminated soil was 10⁷ CFU/g, and the proportion of bacteria reached 99.8%-99.9% of all microbes. This indicated that majority of the crude oil biodegradation was the result of bacterial activity in collaboration with fungi rather than actinomycetes. The microbial activity of uncontaminated soil was higher than that of crude oil-contaminated soil, and microbial activity decreased with increased concentrations of crude oil. This phenomenon can be easily explained by the fact that the microbial metabolic activity had been affected owing to an increase in carbon sources and an imbalance in the soil nutrient ratio followed by an increase in the crude oil concentration in soil. The microbes in both crude oil-contaminated and uncontaminated soil were more likely to use carbon sources such as carbohydrates and polymers on the Biolog plates. Microbes from crude oil-contaminated soil used less of the available carbon sources and showed lower metabolic activity than microbes from the uncontaminated soil. This indicated that soil microbes adapted to the crude oil-contaminated environment by adjusting the microbial community structure, and a correlation was observed between the soil microbial community structure and soil microbial growth. The principal component analyses results revealed a significant difference (P < 0.01) in soil microbial community structure between uncontaminated and crude oil-contaminated soils. The differences mostly related to the use of carbohydrates as the dominant carbon source and then carboxylic acids and amino acids. The variation in the canonical variable (discrete value) increased with increasing soil crude oil content, however, the stability of the soil microbial community structure decreased. This indicated that the crude oil pollutant destroyed the original soil ecological environment. The diversity of microbial community, as indicated by Shannon (H), McIntosh (U), and Simpson (1/D) indices, was significantly different in crude oil-contaminated soil (P < 0.01) compared with uncontaminated soil. H and U values were lower in crude oil-contaminated soil than in uncontaminated soil, and 1/D was higher in crude oil-contaminated soil than in uncontaminated soil. This phenomenon was likely due to the stimulating effect of certain levels of crude oil on the growth of the dominant microbial community. The findings stated above provide a basis for bioremediation of oil-contaminated soil in the Loess Plateau in northern Shaanxi. These results are especially important because they indicated that the soil in the Loess Plateau in northern Shaanxi shows good potential for bioremediation, and crude oil contamination in the soil could be degraded by indigenous microbes with the addition of nitrogen and phosphorus. An additional benefit is that it leads to an improved evaluation of the bioremediation potential of the indigenous microbial consortia.