Mercury is a global pollutant with widespread distribution due to long-range atmospheric transport. Complex biological and chemical interactions prolong the cycling of mercury and affects releases to the environment. Recent global mercury budget estimates indicate that contributions to the atmospheric Hg pool due to natural emissions and re-emissions (i.e. natural terrestrial surfaces re-emitting mercury previously deposited from anthropogenic sources) are equal important to contributions from industrially emitted mercury; two-thirds of this biogenic emission is estimated to be emitted from agricultural soils and the remainder from forest and other shaded-soil landscapes. Several factors, however, complicate the quantification of mercury fluxes over terrestrial systems including scientific uncertainties in mercury cycling processes and microclimatic parameter variability. Although several studies based on micrometeorological approaches have been conducted to determine emission rates of Hg from vegetated and non-vegetated soils, modeling studies about mercury emissions at bare or vegetated soil-interfaces are still limited although necessary for a better understanding of mercury cycling and for the assessment to regulatory perspectives. This paper deals with the application of simple modeling techniques for predicting the flux of atmospheric emission and deposition of gaseous mercury in various terrestrial landscape components. A deterministic environmental simulation model and various statistical models are used. The deterministic model has been developed by the authors and it has been tested and published previously. The model couples a simple air-surface exchange parameterization with algorithms for mercury cycling and transport in surficial soils, allowing, therefore, predictions of cycling, emission and transport fluxes for the various mercury components (elemental mercury, divalent mercury, methyl mercury). Initial soil mercury concentrations are specified in the model, along with wet and dry atmospheric mercury deposition fluxes. The statistical models (parametric and non-parametric models including also neural network models) used in this study have been recently applied by the authors to various airborne pollutants deposited on terrestrial receptors (vegetation, soils). Specifically, the following statistical methods are progressively used in this study: the multiple linear regression model (MLR) and the stepwise multiple linear regression (SMLR); the regression tree method and more specifically the CART (Classification And Regression Tree) technique; the Generalized Additive model (GAM); the locally-weighted smoother model (LOWESS); and the multilayer perceptron model (MLP), the most commonly used artificial neural network model. All models are applied to compare model predictions for air-surface exchange rates of mercury vapor against several sets of contrasting observational data published recently by other authors. In general, the models showed a relatively good agreement between measured and modeled Hg fluxes and demonstrated a good ability to reflect the influence of some basic parameters on Hg emissions from vegetated and non-vegetated soils. The models reflected trends of observational data showing that mercury fluxes are sensitive to the cyclic variations of the various microclimatic parameters; the emission rate of elemental Hg, formed by the reduction of soil divalent mercury, is primarily controlled by changes in solar radiation, soil moisture, temperature and, to a lesser degree, wind conditions. In addition, the sensitivity analyses performed using the environmental model indicated that the degree of sensitivity to model parameters varies between the considered base case landscape systems. These model results suggest that different soils in landscape components result in different processes governing the mercury cycle and air emission fluxes.