Conversion of former agricultural land to grassland and forest ecosystems is a suggested option for mitigation of increased atmospheric CO2. A Sharpsburg prairie loess soil (fine, smectitic, mesic Typic Argiudoll) provided treatments to study the impact of long-term land use on soil organic carbon (SOC) content and composition for a 130-year-old cropped, pasture and forest comparison. The forest and pasture land use significantly retained more SOC, 46% and 25%, respectively, compared with cropped land use, and forest land use increased soil C content by 29% compared with the pasture. Organic C retained in the soils was a function of the soil N content (r=0.98, P<0.001) and the soil carbohydrate (CH) concentration (r=0.96, P<0.001). Statistical analyses found that soil aggregation processes increased as organic C content increased in the forest and pasture soils, but not in the cropped soil. SOC was composed of similar percentages of CHs (49%, 42% and 51%), amino acids (22%, 15% and 18%), lipids (2.3%, 2.3% and 2.9%) and unidentified C (21%, 29% and 27%), but differed for phenolic acids (PAs) (5.7%, 11.6% and 1.0%) for the pasture, forest and cropped soils, respectively. The results suggested that the majority of the surface soil C sequestered in the long-term pasture and forest soils was identified as C of plant origin through the use of CH and PA biomarkers, although the increase in amino sugar concentration of microbial origin indicates a greater increase in microbial inputs in the three subsoils. The practice of permanent pastures and afforestation of agricultural land showed long-term potential for potential mitigation of atmospheric CO2.