We have developed a new high-resolution (1/2° latitude by 2/3° longitude) nested-grid mercury (Hg) simulation over North America employing the GEOS-Chem global chemical transport model. Emissions, chemistry, deposition, and meteorology are self-consistent between the global and nested domains. Compared to the global model (4° latitude by 5° longitude), the nested model shows improved skill at capturing the high spatial and temporal variability of Hg wet deposition over North America observed by the Mercury Deposition Network (MDN) in 2008–2009. The nested simulation resolves features such as land/ocean contrast and higher deposition due to orographic precipitation, and predicts more efficient convective rain scavenging of Hg over the southeast United States. However, the nested model overestimates Hg wet deposition over the Ohio River Valley region (ORV) by 27%. We modify anthropogenic emission speciation profiles in the US EPA National Emission Inventory (NEI) to account for the rapid in-plume reduction of reactive to elemental Hg (IPR simulation). This leads to a decrease in the model bias to +3% over the ORV region. Over the contiguous US, the correlation coefficient (<i>r</i>) between MDN observations and our IPR simulation increases from 0.63 to 0.78. The IPR nested simulation generally reproduces the seasonal cycle in surface concentrations of speciated Hg from the Atmospheric Mercury Network (AMNet) and Canadian Atmospheric Mercury Network (CAMNet). In the IPR simulation, annual mean reactive gaseous and particulate-bound Hg are within 80% and 10% of observations, respectively. In contrast, the simulation with unmodified anthropogenic Hg speciation profiles overestimates these observations by factors of 2 to 4. The nested model shows improved skill at capturing the horizontal variability of Hg observed over California during the ARCTAS aircraft campaign. We find that North American anthropogenic emissions account for 10–22% of Hg wet deposition flux over the US, depending on the anthropogenic emissions speciation profile assumed. The percent contribution can be as high as 60% near large point emission sources in ORV. The contribution for the dry deposition is 13–20%.