[Objective] In order to provide theoretical basis for the evaluation of mountain soil quality under typical vegetation types, land use structure adjustment, soil erosion control and ecological restoration in Southern Fujian Province, we explored the fractal characteristics of soils under different vegetation types in Southern Fujian Province and analyzed the correlations of fractal characteristics with soil particle composition and nutrient content.[Methods] The fractal dimension D values of the soil under three typical vegetation forests in a mountainous region of Southern Fujian Province were calculated using the fractal dimension model of soil particle volume, and their correlations with soil particle composition and soil nutrients were analyzed.[Results] The fractal dimension D values of the soil particle volume, in descending order, were the secondary forest, the Masson pine forest, and the citrus forest respectively. The fractal dimension D value was significantly and positively correlated with the volume percentage of the soil particles of clay and that of powder, and significantly and negatively correlated with the mass percentage of the sand(p<0.05). With regard to the soil pH value, the contents of total nitrogen, total phosphorus, total potassium, hydrolytic nitrogen, available phosphorus, available potassium and organic carbon, their descending order was secondary forest> Masson pine forest > citrus forest; while for soil bulk density, it had a rank as citrus forest > Masson pine forest > the secondary forest. The correlation analysis showed that the fractal dimension D value was significantly and positively correlated with the content of hydrolytic nitrogen in the soil of citrus forest, while it was significantly and negatively correlated with its organic matter. With respect to the Masson pine forest, it had a significant positive correlation with the contents of total phosphorus and total potassium in the soil, and had an extremely significant and positive correlation with the content of hydrolytic nitrogen. At the same time, it had a significant negative correlation with the content of organic carbon. With respect to the secondary forest, it had an extremely significant positive correlation with the content of total phosphorus in the soil(p<0.01), a significant positive correlation with the content of hydrolytic nitrogen, a significant negative correlation with the content of available potassium and organic carbon.[Conclusion] Vegetation types in the southern Fujian mountain area significantly affected the fractal dimension and nutrient contents of the surface soil, and it is possible to use the soil fractal dimension to reflect the soil particle composition and nutrient contents of different vegetation types.