The manufacturing of thin films with structured surfaces by large‐scale rolling has distinct advantages over other techniques, such as lithography, due to scalability. However, it is not well understood or quantified how processing conditions can affect the microstructure at different physical scales. Hence, the objective of this investigation is to develop a validated computational model of the symmetric forward‐roll coating process to understand, predict, and control the morphology of carbon nanotube (CNT)–polydimethylsiloxane (PDMS) pastes. The effects of the thin‐film rheological properties and the roller gap on the ribbing behavior are investigated and a ribbing instability prediction model is formulated from experimental measurements and computational predictions. The CNT–PDMS thin‐film system is modeled by a nonlinear implicit dynamic finite‐element method that accounts for ribbing instabilities, large displacements, rolling contact, and material viscoelasticity. Dynamic mechanical analysis is used to obtain the viscoelastic properties of the CNT–PDMS paste for various CNT weight distributions. Furthermore, a Morris sensitivity analysis is conducted to obtain insights on the dominant predicted characteristics pertaining to the ribbing microstructure. Based on the sensitivity analysis, a critical ribbing aspect ratio is identified for the CNT–PDMS system corresponding to a critical roller gap.