International audienceIn recent decades, two of the main management tools used to stem biodiversity erosion have been biodiversity monitoring and the conservation of natural areas. However, socio-economic pressure means that it is not usually possible to preserve the entire landscape, and so the rational prioritisation of sites has become a crucial issue. In this context, and because floodplains are one of the most threatened ecosystems, we propose a statistical strategy for evaluating conservation value, and used it to prioritise 46 waterbodies in the Loire floodplain (France). We began by determining a synthetic conservation index of fish communities (Q) for each waterbody. This synthetic index includes a conservation status index, an origin index, a rarity index and a richness index. We divided the waterbodies into 6 clusters with distinct structures of the basic indices. One of these clusters, with high Q median value, indicated that 4 waterbodies are important for fish biodiversity conservation. Conversely, two clusters with low Q median values included 11 waterbodies where restoration is called for. The results picked out high connectivity levels and low abundance of aquatic vegetation as the two main environmental characteristics of waterbodies with high conservation value. In addition, assessing the biodiversity and conservation value of territories using our multi-index approach plus an a posteriori hierarchical classification methodology reveals two major interests: (i) a possible geographical extension and (ii) a multi-taxa adaptation