Carbon sequestration by soils in different production systems contributes greatly to the reduction of greenhouse gases. The objective of this study was to quantify the carbon stored in four land uses at different soil depths. To this end, a 22 factorial experiment in complete randomized block design (CRBD) was carried out. The factor A: land uses (natural pastures, shelterbelts, fodder banks, and potato crop) and the factor B: two soil depths (30 and 60cm), with three replications. . As a result, statistical differences were found among soil uses (p>0.0573) and between depths of 30 and 60cm (p<0.0061). However, no statistically significant differences were found in the interaction land-use and depth (P > 0.0659). The fodder bank presented a higher organic carbon content (139.85tC.ha-1) at 60cm depth and the potato monoculture (63.32tC.ha-1) at 30cm depth while, at both depths, natural pasture reported lower values (54.45 and 60.02tC.ha-1). Hence, the importance of productive systems to accumulate more carbon at greater depths of soil (60cm) compared to lower depths (30cm), which may be linked to agricultural opperations made on the soi surface, generating carbon leakage.