Coffee agroforestry systems (CAFS) sustain the livelihoods of many people globally at the same time as providing important ecosystem services such as carbon sequestration that help mitigate climate change. These systems vary in their composition (especially density and species of shade tree) and management. Changes made to enhance their productivity will affect their climate change mitigation potential. With growing food demand and diminishing availability of agricultural land due to global population growth, as well as an increasing threat from global climate change the trade-offs between the socio-economic and net carbon sequestration performance in CAFS are important. The carbon sequestration and socio-economic performance of a range of CAFS varying in composition and management were assessed in Costa Rica and Nicaragua. Measurements and modelled estimates were made of (i) greenhouse gas emissions (GHGs) from coffee cultivation (the carbon footprint (CF)), (ii) carbon sequestration potential into above-ground biomass and soil organic stocks and (iii) socio-economic performance (productivity and profitability), and their trade-offs analysed. The effects of agronomic management (conventional versus organic) and shade type (ranging from timber trees to full sun) on the CF of two long-standing CAFS experiments in Costa Rica Nicaragua demonstrated that management is the best predictor of the CF whereas shade type has a minor effect. The greatest contributor to the overall CF was N2O emissions from the input of N in applied organic and inorganic fertilisers. Shade systems with high levels of N input from leguminous tree pruning had the highest CF. Total soil organic carbon (SOC) decreased over the first nine years of coffee bush and shade tree establishment in these experiments, although this differed amongst soil layers. Organically managed systems tended to have an increase in SOC in the top 10 cm of soil, though organic and conventional systems had similar (larger) decreases in SOC in deeper soil. Shade type and above-ground biomass had a smaller effect on SOC. Comparison of the CF of these experimental CAFS treatments with their C sequestration potential showed that increases in GHG emissions from production intensification can be compensated for or even outweighed by the increase in C sequestration into above-ground biomass, especially for shaded systems. However, if less productive, lower intensity CAFS are extended onto an area of currently forested land in order to compensate for the shortfall in profitability (compared with higher-intensity, higher-yielding systems), this land-use change causes additional GHG emissions from deforestation. This results in net GHG emissions for the whole system for the majority of shade types tested. Evaluation of the C and socio-economic performance of coffee farms in the regions around the two experimental sites showed that due to the huge variation amongst CAFS there is no single strategy for climate change mitigation that could successfully be applied across the range of farms. Instead it will be necessary to carry out accurate and site-specific farm assessments to inform advice and decisions on system improvement tailored to the needs of individual farms and environmental settings. The findings of this research suggest that there is a place in the C market for CAFS, however their design and management will determine the overall net benefits that can be achieved.