Polycyclic aromatic hydrocarbons (PAHs) are an important group of persistent organic pollutants that are widespread in environment. With rapid growth of population, sustained industrial development and urbanization, PAHs have become a growing concern and received much attention. The reason is that PAHs have toxicity to humans and their possible harmful effect on animals, plants and microbes in aquatic and terrestrial ecosystems. In order to reveal the responses of tree species to PAHs and to examine the remediation methods of diesel-oil contaminated soils using higher plants, a pot experiment was conducted to compare degradation of PAH concentrations in diesel-oil contaminated soils under four tree species, and to investigate how soil microbial flora make contributions on such degradation. One-year old seedlings of four tree species, Cinnamomum camphora, Magnolia grandiflora, Koelreuteria bipinnat, and Liriodendron chinense were selected and planted in PAH contaminated soils with different diesel-oil (0g/kg (control), 2g/kg (L1), 10g/kg (L2) and 50g/kg soil (L3)). The composition, abundance and dynamics of soil microbes in the four groups of soils were measured to provide theoretic evidences for assessing and selecting tree species for remediation of PAHs contamination. The results showed that (1) bacteria were the most dominant group in soil microbial flora in the four tree species, and accounted for 69.70%-93.87% of the total microorganism. The percentage of bacteria usually increased with increasing of diesel-oil concentration levels. Compared with in control treatments, the proportion of bacteria in soil microorganisms in PAH treatments were inhibited by diesel-oil concentrations, except in soil of Koelreuteria bipinnat. Actinomyces accounted for 2.71%-25.29% of total microbeorganisms, and the percentage decreased with the increases of diesel-oil concentrations in all treatments. For Koelreuteria bipinnat species, the percentages of Actinomyces in total soil microbial flora were smaller in contaminated treatments than in control. But the situation was opposite for Cinnamomum camphora species. The proportion of fungi in total soil microorganisms was 0.88%-5.83%, and the values were relatively promoted in low and medium diesel-oil concentration treated soils. (2) The microorganisms in the studied tree species soils exhibited different seasonal patterns to response to PAHs treatments. In Koelreuteria bipinnat soils, the number of total microbes was higher in contaminated treatments than in control; In Cinnamomum camphora soils, only at the early stage were the total microbes high in polluted treatments than in control. For Magnolia grandiflora species, the abundance of total microorganisms was higher in all diesel-oil treated soils than in control soils for the studied period, except in the month of April, when the number of microbes was higher in control than in other treatment soils. The total microbes were lower in polluted treatments than in control in Liriodendron chinense soils in April. But in other months, all treatment soils had higher microbes than in control except in L1 treatment. (3) In control soils, the quantity of total microbes gradually increased at the early stage of the study (from October, 2009), reaching the peaking value in April, 2010, and then decreased until in October, 2010. PAHs contaminations disturbed such dynamic patterns of microbial abundance. The maximum values of total microorganisms in were appeared in January or July in the treatments. It was found that fungi played an important role in controlling the degradation of PAHs in this study. One year later of the treatment, PAHs contents were similar in L1, L2 treatments' soils and in control soils for all four tree species plots. PAHs contents in L3 treatment's soils of each tree species were in the order of Liriodendron chinense > Koelreuteria bipinnat> Magnolia grandiflora>Cinnamomum camphora.