The fifth framework EU project BORIS (Bioavailability Of Radionuclides In Soils: role of biological components and resulting improvement of prediction models) had three scientific objectives. The first was to improve understanding of the mechanisms governing the transfer of radionuclides to plants. The second was to improve existing predictive models of radionuclide interaction with soils by incorporating the knowledge acquired from the experimental results. The last and third objective was to extract from the experimental results scientific basis for the development of bioremediation methods of radionuclide contaminated soils and to apprehend the role of additional non-radioactive pollutants on radionuclide bioavailability. This paper is focused on the second objective. The purpose of the BORIS assessment model is to describe the behaviour of radionuclides in the soil-plant system with the aim of making predictions of the time dynamics of the bioavailability of radionuclides in soil and the radionuclides concentrations in plants. To be useful, the assessment model should be simple and use only a few parameters, which are commonly available or possible to measure for different sites. The model shall take into account, as much as possible, the results of the experimental studies and the mechanistic models developed in the BORIS project. The adopted approach was to introduce in the assessment model a quantitative relationship between bioavailability of the radionuclides in soil and the soil properties. To do this an operational definition of bioavailability was proposed. Here, operational means experimentally measurable, directly or indirectly, and that the bioavailability can be translated into a mathematical expression. This paper describes the reasoning behind the chosen definition of bioavailability and how it was used in the assessment model.