¹³⁷Cs is a radionuclide with a half-life of 30 years that is commonly found in soils after nuclear fallout due to nuclear incidents or atmospheric nuclear weapon testing. Due to their properties of accumulation and retention, soils are key compartments for the transfer of contaminants such as ¹³⁷Cs in the trophic chain. Ingestion of contaminated agricultural products being one of the main component of human exposure, it is essential to be able to predict the fate of ¹³⁷Cs throughout the soil-plant continuum. The contaminant mobility into the soil, its transfer to the plant and its final distribution between all components are generally described by simple models (equilibrium-based, linear distributions). These models are operational but are not able to account for the variability of soils and plants encountered. Bioavailability is function of both soil physico-chemical characteristics, that impact the environmental availability, and plant physiology which determines the uptake rate and accumulation. The aim of this work is to highlight the preponderant factors controlling the ¹³⁷Cs bioavailability in the soil-solution-plant continuum by using a model that account for both soil and plant characteristics. The proposed mechanistic model is based on thermodynamic reactions describing the interactions of Cs with the different soil reactive components, coupled with a physiological model of root absorption. Series of experiments were conducted to produce a contrasted data set of ¹³⁷Cs soil to plant transfer. For those experiments, 2 different plants with contrasted Cs uptake capacities (Millet, Mustard) and 3 different soils with varying texture and mineralogy have been studied. Three weeks exposure studies were conducted with the RHIZOtest® which is a normative device to assess the bioavailability of contaminants in soil. They were completed with batch experiments aiming at characterizing the environmental availability of Cs in soils. A large range of ¹³⁷Cs soil to plant transfer rates was measured for the different soil/pant combinations. For example a contrasted bioavailability of ¹³⁷Cs was observed, with the same plant accumulating 10% to 40% of total Cs’s stock depending on the soils. We also observed that during the time of the experiment the plant had absorbed most of the estimated environmental available Cs. Modelling those experiments allowed us to highlight the main soil and plant properties that have a great impact on the contaminant mobility. For example, plant physiological factors were the main driver of bioavailability in condition where environmental availability was not limited (e.g. sandy soils), whereas their roles were reduced in soils with high ¹³⁷Cs sorption capacity. Such a model may help to reduce uncertainties in the prediction of ¹³⁷Cs transfer to plants in environmental risk assessment, with a great potential to cover a large range of soils and plants.