Thoracic ossification of ligament flavum (TOLF) is a pathological process of heterotopic bone formation from ligamentum flavum tissue, which is the most common cause of thoracic spinal stenosis and thoracic myelopathy. Imaging examination is the predominant measure for localization and qualitative diagnosis for TOLF. Various TOLF classification systems have been reported currently, based on its morphology, distribution, configuration of compressed spinal cord or its association with peripheral tissue structures. However, there has been no unified classification due to confusing clinical applications. Therefore, we conducted a review to summarize TOLF classification systems and to evaluate the reliability of these classification systems and the diagnostic accuracy of various imaging modalities. As demonstrated in literatures, the diagnostic accuracy of radiography and the reliability of related classification were low. CT-based Sato classification (lateral, extended, enlarged, fused, tuberous) and Kuh classification (unilateral, bilateral, bridging), MRI-based Kuh classification (round, beak) and Chen Zhongqiang classification (isolated, skipping, continuous), Muthukumar classification combined with dural ossification (Tram track sign, Comma sign) elaborated ossification characteristics from different angles and dimensions. These classification systems were clinically significant in pathogenesis implication, risk assessment, treatment guidance and prognosis judgement, which showed high diagnostic accuracy and reliability. Combining multi-dimension and multi-level parameters of CT and MRI will play an important role in the diagnosis, treatment and new classification research of TOLF.