Different density functional theory (DFT) approaches were tested for the computation of 1H and 13C nuclear magnetic resonance (NMR) chemical shifts in monosubstituted ferrocenes. The results were evaluated against experimental values. Generally, the conductor‐like polarizable continuum model and cc‐pVTZ basis set are recommended. The geometries providing the best accuracies are B3LYP‐optimized for 1H and M06‐L‐optimized for 13C. Functional rankings at these geometries are: TPSSh > M06‐L > CAM‐B3LYP > B3LYP > PBE0 > M06 (the first one is the most accurate) for 1H NMR computations and M06 > M06‐L > PBE0 > TPSSh > B3LYP > CAM‐B3LYP for 13C. The most accurate functionals have root‐mean‐square deviations of 0.08 ppm (1H, TPSSh) and 3.97 ppm (13C, M06) and showed similar accuracy for a set of disubstituted ferrocenes and decamethylferrocene. The utilization of Jensen's pcSseg‐2 basis set improves the results for 1H but worsens the results for 13C. The linear scaling is generally not recommended. The errors can be minimized using an appropriate method for a given nucleus, so the DFT‐assisted signal assignment is possible for substituted ferrocenes.