There is actually a surge for reintroducing biodiversity in agricultural systems in order to reduce chemical inputs, suppress pests, and close biogeochemical cycles (Malézieux et al., 2009). The use of cover-crops is a promising way to reintroduce biodiversity into the fields. Cover-crops have the potential to decrease chemical use against weeds (by competition) and pests (by increasing in natural enemies). To decrease herbicide use, suitable plants must be able to grow in appropriate conditions, to do not compete the cultivated crop for nutrients or water, but should compete weeds for light and space. There is a trade-off between these objectives. Banana is a semi-perennial crop, each plant develops at its own rhythm leading to an unsynchronized plant population in three years; canopy and nitrogen demand of the crop follow this unsynchronized pattern (Tixier et al., 2004). Banana cropping systems remain based on bare soil management and a large amount of herbicides is used. In tropical environment, the growth of weeds and cover-crop is complex because it is not constrained by seasons; a constant growth is possible due to relatively constant climate. In these conditions, variation in radiation due to canopy closure is one of the major drivers of their growth. First, we present the evaluation of the impact of the cover-crop on the yield in the case of two graminaceous cover-crops. Secondly, we present results of the evaluation of 11 species of potential cover-crop using measures under different levels of shade. We used a simulation model to contextualize their growth and conclude on the most appropriated species to intercrop with banana.