Background, aim, and scope Land-use change can significantly influence carbon (C) storage and fluxes in terrestrial ecosystems. Soil-plant systems can act as sinks or sources of atmospheric CO 2 depending on formation and decomposition rates of soil organic matter. Therefore, changes in tropical soil C pools could have significant impacts on the global C cycle. This study aims to evaluate the impacts of long-term sugarcane cultivation on soil aggregation and organic matter, and to quantify temporal dynamics of soil organic matter in cultivated sugarcane plantation soils previously under a tropical natural secondary forest. Materials and methods The soil in the study area was an Ultisol rich in Fe oxides. Soil samples were taken from sugarcane land converted from natural secondary forest 35 years (SC35) and 56 years (SC56) previously. Soil from an adjacent, continuous natural secondary forest (CNSF) was also taken for comparison. Soils were taken from four depths to 1 m, and fractionated by size (>250 μm, 53-250 μm, and <53 μm) and density (>53 μm). Each soil fraction was analyzed for organic C concentration and the 13 C isotopic signature δ 13 C. Results and discussion Compared with CNSF, SC35 and SC56 soils were characterized by higher proportions of microaggregate (53-250 μm) and silt&clay (<53 μm) fractions. Soil δ 13 C values indicated that sugarcane (a C 4 plant) cultivation resulted in sequestration of new C (sugarcane derived), but significant loss of old C (native forest derived) in the soil organic matter fractions. Isotopic analysis indicated that sugarcane-derived biomass contributed more than 33% and 25% of the total organic C in the SC35 and SC56 soils, respectively. After 56 years of sugarcane cultivation, organic C concentrations in the total soil and in each fraction were significantly lower than in the CNSF soil, with a reduction of greater than 60%. Although organic C concentration in the SC35 soil was lower than in the CNSF soil, the difference was not statistically significant. Sugarcane cultivation caused a loss of organic C in the upper soil layers through both enhanced microbial decomposition and downward translocation in the soil profile. Conclusions When converting native forest land to sugarcane cultivation, soil organic C reduction will continue for a very long period (e.g., over several decades) before a new equilibrium is established. Despite large losses of total organic C, particularly in the SC56 soil, decades of continuous sugarcane cultivation resulted in only a relatively small though significant reduction in the macroaggregate fraction. This implies that Fe oxides rather than organic matter are the dominant binding agents for macroaggregate formation in this Fe-rich soil. Subsequently, the persistence of Fe oxide-bound macroaggregates may have prevented soil organic matter from rapid decomposition when native forest soil was cultivated.