The multiplicity of above-ground biomass uses is a major issue regarding conservation agriculture (CA) systems (Erenstein, 2003; Giller et al, 2009). We have simulated the biomass potential uses in the two CA cropping systems usually applied in the Alaotra Lake region of Madagascar for producing rainfed rice. The first is a two year rotation between maize intercropped with Dolichos lablab (MD) followed by rice. The second is a three year rotation system of maize intercropped with Stylosanthes guianensis, followed by one year of S. guianensis (SG) alone, ultimately followed by rice. The first rotation is supposed to produce more grain than the second one because of the presence every year of a cereal production. We compared these two systems regarding major tradeoffs between different uses linked with three different objectives: i) crop production, ii) animal production and iii) other ecosystem services such as erosion control. We focused on the use of the biomass produced the year preceding our main crop of rice, i.e. maize+ D. lablab or S. guianensis alone and evaluated the direct, or indirect consequences of its removal on the three objectives mentioned above. The purpose of this communication is to see which conceptual model can be useful to study these trade-offs and to give some first values and range for some outputs of the two cropping systems studied.