The measurement of diffusive gradients in thin films (DGT) has been proposed as a surrogate for metal uptake by plants. A small-scale experiment was performed to test the predictive capacity of the DGT method with respect to uranium and heavy metal availability and uptake by ryegrass. Correlation analysis was performed to compare the results obtained with the DGT device with more conventional bioavailability indices: concentration in pore water or in selective extractants. Six soils with different histories of uranium contamination and with distinct soil characteristics were screened. For uranium, the bioavailability indices evaluated were highly correlated, indicating that uranium pools assessed were at least partially comparable. For heavy metals, bioavailability indices were hardly correlated. Uranium concentration in the pore water was a better predictor for uranium uptake than amount recovered following selective extraction. The concentration measured with DGT (CDGT) was also highly correlated with plant uptake. However, the significance level was sensitive to the value of the diffusion coefficient (pH dependent or not) used to calculate CDGT. For the heavy metals screened, only in the case of Cd was uptake predicted by CDGT.