Radioecological models used to make predictions of the radionuclide activity concentrations in human foodstuffs must be sufficiently robust and fit for purpose with uncertainties reduced where practicable. The CONFIDENCE project had a work package with the objective to improve the capabilities of radioecological models and this paper presents the key findings of this work. Recommendations for future radioecological studies/model developments are made based on the findings of the work conducted and consultations with end-users.