[Objective] The effects of changes in aboveground litter, underground roots, and mycorrhizal inputs on soil microaggregates of restored red soil forestland were analyzed in order to provide a basis for soil function reconstruction and understanding the ecosystem carbon cycle after forest restoration of degraded lands. [Methods] The study was conducted in a pure Schima superba forest, a typical broad-leaved forest recovered from eroded and degraded red soil in a subtropical region. Five input treatments were established: no litter (CT), mycorrhiza (M), root+mycorrhiza (RM), aboveground+underground litter (LRM), and double aboveground+underground litter (DLRM). The composition of soil microaggregates, organic carbon, and their chemically bound forms were analyzed. [Results] There were no significant differences in soil microaggregate mass percentage, organic carbon, Ca-SOC, Fe (Al)-SOC, and Ca-SOC/SOC among the different treatments (p＞0.05). Compared with CT, LRM reduced Fe (Al)-SOC/SOC of microaggregates of size 20—50 μm and 50—200 μm by 40.06% and 46.67%, respectively (p＜0.05). Soil microaggregate mass percentage, organic carbon, Ca-SOC, and Fe (Al)-SOC decreased as particle size increased, and organic carbon and bound organic carbon tended to be enriched in smaller particle size groups (＜20 μm). The content of Ca-SOC in microaggregates (0.55—1.28 g/kg) was much lower than observed for Fe (Al)-SOC (6.88—13.34 g/kg), but its variation range in different particle sizes was greater than that of Fe (Al)-SOC. The percentage of Ca-SOC/SOC (1.54%—3.44%) was also less than the percentage of Fe (Al)-SOC/SOC (16.75%—42.54%). Microaggregate mass percentage, organic carbon, Ca-SOC, and Fe (Al)-SOC were positively correlated (r=0.497—0.757, p＜0.01). [Conclusion] Soil microaggregates and their organic carbon in a restored S. superba forest were affected by particle size. Both quantities responded to short-term inputs of aboveground litter, underground roots, and mycorrhiza, but did not reach a significant level, and therefore need to be studied over a longer time period.