Magnetic resonance imaging (MRI) and chemosensory event-related potentials (ERPs) are important methods to evaluate olfactory function, but there is lack of study to explore the application of MRI and chemosensory ERPs in the patients with traumatic anosmia. The data of 26 post-traumatic anosmic patients and 21 healthy controls were retrospectively surveyed; olfaction and olfactory pathway of all participants were measured clinically using the T&T olfactometer, the Sniffin' Sticks, chemosensory ERPs and MRI. All patients were anosmic based on complaints and clinical examinations. In five patients, the olfactory bulb volume was significantly lower than control group. In 18 patients, the olfactory sulcus (OS) depth was similar to control group, but all the participants had a deeper right OS (right = 7.79 +/- A 1.31, left = 7.06 +/- A 1.44, p < 0.01). Olfactory ERPs (oERPs) could be evoked in 17 patients, but these signals showed longer latencies and lower amplitude than controls in the N1 (latency p < 0.05, amplitude p < 0.01) and P2 (latency p < 0.01, amplitude p < 0.05) waves. Nine traumatic anosmic patients had no identifiable oERPs; most of them had olfactory center injury. Trigeminal ERPs were detected in all anosmic patients and controls; patients had longer latencies for N1 (p < 0.05) and P2 (p < 0.05) waves, while there was no similar change in amplitude. Older subjects had smaller OB volume and OS depth. Closed head injury could induce anosmia; the severity extent, injury site and subsequent consciousness are related to the olfaction. oERP is the gold standard for olfactory subjective examination; MRI could indicate the lesions on the olfactory pathway and reflect the possibility of detectable oERPs.