There has been renewed interest in the application of models to the transport of non-point source pollutants. However, very little work has been done to evaluate the performance of a functional transient-state model for the transport of a reactive solute over an extensive study period. This research consists of mathematical modeling to simulate water flow, boron and selenium transport through soil in tile-drained croplands.For Boron part a mathematical model was developed to simulate non-conservative boron transport. The dynamic two-dimensional finite element model simulates water flow and boron transport in saturated-unsaturated soil system, including boron sorption and boron uptake by root-water extraction. Two different models have been employed for the sorption of boron. Similarly, for selenium part a finite element model is developed to simulate species of selenium transport in two dimensions in saturated/unsaturated zones. The model considers water, selenate, selenite and selenomethionine uptake by plants. It also considers oxidation/reduction, volatilization, and chemical and biological transformations of selenate, selenite, and selenomethionine. Comparison of boron transport model results with observed data is satisfactory. The model employed with Langmuir isotherm was found to give slightly better simulation results when compared with the model employed with Freundlich. The sensitivity analysis results indicate that the irrigation scheduling and the irrigation water quality are very important parameters for boron accumulation in the soil. Also the adsorption isotherm parameters, which reflect us the soil properties, are found to be important for the boron movement in the soil. Comparison of selenium transport model with observed data is not quite satisfactory in accuracy when compared with the model for boron transport. This may be the result of the complexity of the mechanisms affecting the selenium transport in soil. There are too many parameters, and due to the errors depending on the parameters, the total error for the estimation of the total selenium increases.