Agroforestry can raise carbon (C) stocks of agricultural systems, and such increases can potentially be sold as CO emission offsets. We assembled information on the biophysical, technical, economic, and practical potential of agroforestry to sequester C for the West African Sahel, East Africa, and Southern Africa. Agroforestry systems (AFS) such as parklands, live fences, and homegardens had substantial C stocks, but only accumulated 0.2–0.8 Mg C ha−1 year−1. Rotational woodlots (2.2–5.8 Mg C ha−1 year−1) and possibly improved fallows in Southern Africa sequestered C relatively faster, but only during the fallow phases. Data on soil C are scarce because most studies only compared soil C under different land uses, which provides limited (and sometimes unreliable) information on sequestration rates. Comparing results from different studies is difficult, because no standard protocols exist. Few studies have evaluated the economic potential of agroforestry to sequester C. However, at prices of Dollar 10 per Mg CO -eq or less, the value of stored 2 C in most systems would be less than Dollar 30 ha−1 year−1, which is a small fraction of annual farm revenue and it needs to cover all transaction measurement reporting and verification costs. Practical constraints to C sequestration (CS) such as land tenure, policy issues, and the opportunity costs incurred by possibly foregoing more profitable land management options have not been fully explored for Africa. For evaluating the challenges and opportunities involved in CS by smallholder farmers, comprehensive studies are needed that explore all C and non-C costs and benefits of agroforestry activities.