The objective of this experimental investigation was to determine the geometric simulation fidelity required to accurately model the aerodynamics of a horn-ice accretion in a wind tunnel. A casting and a 2-D smooth simulation with variable horn geometry were constructed to model a horn-ice accretion on a NACA 0012 airfoil. Several simulations of differing fidelity, including a casting, were constructed to model a horn-ice accretion on a NACA 23012 airfoil. Aerodynamic testing was performed in the University of Illinois 3 x 4 ft wind tunnel at a Reynolds number of 1.8 x 10⁶ and a Mach number of 0.18. Minor changes to the upper-horn geometry of the NACA 0012 2-D smooth simulation were found to have notable impacts on drag and maximum lift. Therefore, spanwise variations in the ice accretion geometry must be carefully examined so that an appropriate cross section can be chosen from which to generate a tracing for a 2-D simulation. Such a 2-D smooth simulation, as was constructed for the NACA 23012 airfoil, can model maximum lift to within 1% of that of the casting. This type of simulation can also provide an estimate of drag that is within the uncertainty of the casting due to spanwise variation, although it does not reproduce three dimensionality in the iced-airfoil flowfield.