Assessment of the impact of naturally occurring radionuclides (NOR) within terrestrial biotopes of uraniferous territories is a sound issue. Among others, it requires deep understanding of NOR transfer within the soil-water-microorganisms/plants continuum. In this study, part of RADONORM (european-funded) and Needs-Environment INSPECT projects, we combined in-situ and lab characterization of samples taken from a small wetland downstream a former U-mine (Rophin site, France). Soil cores and co-located plants (Caltha palustris, Scirpus sylvaticus) were sampled in 2 zones, where a layer of U-rich materials originated from decantation pound leakage is present or not (control) in the root-colonized horizons. Samples were characterized for U and Th contents. Environmental availability and bioavailability were assessed in some soil layers respectively by successive batch desorption and by the ISO normalized RHIZOtest with ryegrass as test plant. U has an heterogenous distribution within the soil profile with higher concentration in the U-rich materials layer. In situ-U Transfer Factors ([U]leaves/[U]soil) were not different between the 2 zones despite contrasted soil and plant concentrations. For Th, plant concentrations and TF values were lower in the control zone than in the impacted one despite similar concentrations in soil. Environmental availability of U (0.3 to 3%) and Th (<0.1%) was low in all soil layers, with higher values in organic surface horizons of both zones for Th. These values differ from those of bioavailable U and Th assessed by the RHIZOtest experiments. The latter are lower than the environmental available ones for U in the mine layer and in organic layers for Th, whereas the opposite is observed for U in both horizons of the control zone and for Th in the deeper horizons of both zones. Results highlight the complex link between element speciation, solution/root relationships and transfer intensity and illustrate the remaining challenge to link availability/bioavailability experiments.