This paper describes a dynamic compartment model for evaluating the tritium level in agricultural plants after a short-term exposure to HTO vapor and its comparison with experimental results to test the predictive accuracy of the model. The model uses a time-dependent growth equation of a plant so that it can predict the contamination level of tritium depending on the stage of the growth of the plant, which is a major difference from some other compartment models using a constant crop yield. The model is able to calculate the time variable concentrations of the compartments representing the atmosphere, soil, and plants of four categories including leafy vegetables, root vegetables, grains, and tuber plants. Experimental results include the tissue free water tritium (TFWT) and the organically bound tritium (OBT) concentration of rice, soybean, cabbage, and radish exposed to HTO vapor for 1 h in the daytime at different growth stages. The model predictions showed that the model could simulate well not only the time-dependent tritium concentration of the plants but also the effect of the growth stage of the plant at the exposure time. Comparison of the model predictions with the experimental results suggested that the model could predict reasonably well the observed TFWT and OBT concentrations of the plants considered.