Modeled and measured bi-directional fluxes (BDFs) of ammonia (NH₃) were compared over fertilized soybean and corn canopies for three intensive sampling periods: the first, during the summer of 2002 in Warsaw, North Carolina (NC), USA; and the second and third during the summer of 2007 in Lillington, NC. For the first and the third experimental periods, the BDF model produced reasonable diurnal flux patterns. The model also produced correct flux directions (emission and dry deposition) and magnitudes under dry and wet canopy conditions and during day and nighttime for these two periods. However, the model fails to produce the observed very high upward fluxes from the second sampling period due to the fertilization application (and thus being much higher soil emission potentials in the field than the default model values), although this can be improved by adjusting model input of soil emission potentials. Model-measurement comparison results suggest that the model is likely capable for improving long-term or regional scale ammonia predictions if implemented in chemical transport models replace the traditional dry deposition models, although modifications are needed when applying to specific situations.