We quantify uncertainties in the global mass balance of mercury and identify research priorities to reduce these uncertainties. This is accomplished by developing a new spatially resolved global multimedia model (WorM3) that quantitatively describes the fate of mercury at a process level, and conducting an uncertainty analysis on its unit-world variant which computes similar global estimates. In our modeling approach, all mass transfer processes and reactions in ocean water, soil and vegetation, are represented as pseudo-first order; reactions in air are represented using the ratios of observed concentrations of mercury species. We use Monte Carlo analysis to estimate uncertainties in the unit-world modeled global mass balance of mercury and quantitatively identify the input parameters which contribute most to these uncertainties. A key finding is that uncertainties in input parameters that describe the rates of reduction and oxidation reactions in surface ocean contribute more than uncertainties in anthropogenic emissions to the total uncertainties in atmospheric concentration and depositional fluxes of mercury. More research should therefore be targeted toward understanding of these oceanic processes.