This paper presents an approach for combined biomass inventories in the scope of future REDD regimes. The focus is set on a sound and reliable method for measuring and monitoring the current state of carbon stocks and their changes over time. A reliable framework for measuring, reporting and verification is urgently needed to ensure the integrity and credibility of REDD efforts in general and REDD in the post - 2012 agreement which is assumed to be approved at COP 16 in Mexico in December 2010. The proposed approach was developed and successfully implemented in Madagascar within a multi - institutional REDD project, i.e., REDD - FORECA. It combines a multi - temporal remote sensing approach incorporating satellite sensors from medium to very high resolution with a terrestrial cluster sampling design, which proved to be operational for the whole spectrum from highly fragmented to pristine forest areas. This combination was implemented by a multi - phase sampling approach. The inventory is designed for the prerequisites of a continuous forest inventory to facilitate the quantification of possible CO2 reductions over time. The first field - assessments were accomplished in 2007 and 2008, and resulted in estimates of aboveground biomass on single tree level. Statistical upscaling procedures were utilised to aggregate these estimates on several levels. The results of the introduced methodology are presented and discussed.