The importance of agroforestry systems as carbon sinks has recently been recognized due to the need of climate change mitigation. The objective of this study was to compare the carbon content in living biomass, soil (0-10, 10-20, 20-30 cm in depth), dead organic matter between a set of non-agroforestry and agroforestry prototypes in Chiapas, Mexico where the carbon sequestration programme called Scolel'te has been carried out. The prototypes compared were: traditional maize (rotational prototype with pioneer native trees evaluated in the crop period), Taungya (maize with timber trees), improved fallow, traditional fallow (the last three rotational prototypes in the crop-free period), Inga-shade-organic coffee, polyculture-shade organic coffee, polyculturenon-organic coffee, pasture without trees, pasture with live fences, and pasture with scattered trees. Taungya and improved fallow were designed agroforestry prototypes, while the others were reproduced traditional systems. Seventy-nine plots were selected in three agro-climatic zones. Carbon in living biomass, dead biomass, and soil organic matter was measured in each plot. Results showed that carbon in living biomass and dead organic matter were different according to prototype; while soil organic carbon and total carbon were influenced mostly by the agro-climatic zone (P \ 0.01). Carbon density in the high tropical agro-climatic zone (1,000 m) was higher compared to the intermediate and low tropical agro-climatic zones (600 and 200 m, respectively, P \ 0.01). All the systems contained more carbon than traditional maize and pastures without trees. Silvopastoral systems, improved fallow, Taungya and coffee systems (especially polyculture-shade coffee and organic coffee) have the potential to sequester carbon via growing trees. Agroforestry systems could also contribute to carbon sequestration and reducing emissions when burning is avoided. The potential of organic coffee to maintain carbon in soil and to reduce emissions from deforestation and ecosystem degradation (REDD) is discussed.