Since ice formation on surfaces at subzero temperatures leads to accidents, increased equipment maintenance costs, and reduced performance, multiple strategies, including superhydrophobic surfaces and coatings, have been explored as means to reduce ice adhesion to solid surfaces. Previous work has correlated the effect of topography of regularly patterned superhydrophobic surfaces with ice adhesion. This work, however, investigated the effect of filtered topography on ice adhesion for random superhydrophobic surfaces. The ice adhesion behavior of superhydrophobic composite coatings, prepared from a mixture of silica nanoparticles and polymer binder and sprayed on glass slides, was determined using a shear strength measurement. The ice adhesion significantly decreased with an increase in particle content up to 40 wt.%, after which the ice adhesion became nearly constant. The present study focuses on the use of a novel filtering method for coating topography evaluation which isolated the asperities contributing to the interface from the roughness profile in the superhydrophobic coating. It showed that the ice adhesion correlated with the filtered asperity height and spacing for these random hydrophobic surfaces. Higher particle contents led to larger asperity distances, smaller solid fractions, and lower ice adhesion. The results and conclusions are based on a static ice adhesion test using still water. In this work, it is demonstrated that ice adhesion can be predicted based on the solid–water–air interface, a correlation that could guide future superhydrophobic coating fabrication to create surfaces with greater reduction in ice adhesion.