The pulse labeling technique was used to label ¹³C-enriched wood residue of Cunninghamia lanceolata sapling, which was then mixed, separately, with soil samples collected from the topsoil layers of a citrus orchard (Citrus aurantium ), a fir plantation (Cunninghamia lanceolata), a secondary forest of Castanopsis carlesii, and a natural forest of Altingia gracilipes). To trace dynamics of fresh carbon in soil aggregates, the mistures were incubated at 20 ℃ for 90 d and then analyzed for distribution of ¹³C-labeled wood residue carbon in the three fractions of soil aggregates, i.e. >250 μm, 53~250 μm and <53 μm. Results show that in the soil samples amended with 13C-labeled wood residue, δ¹³C value increased in all the fractions of soil aggregates, especially in the soil samples of the orchard. In all the soil samples, regardless of land use, δ¹³C value was the lowest in the <53 μm fraction of soil aggregates, and the highest in the 53~250 μm fraction. Once extraneous carbon was amended into the soil samples, the content of fresh carbon therein increased instantly. And the content of fresh carbon and δ¹³C value was higher in the root than in the leaf, as the ratio of C/N is higher in the former. In terms of content of fresh carbon, the three fractions of soil aggregates displayed an order of <53 μm > 53~250 μm > >250 μm, which indicates that a short term of incubation facilitates accumulation of fresh carbon in micro soil aggregates.