To investigate the distribution of traditional Chinese medicine (TCM) Syndromes of diabetic retinopathy and explore the correlation between various symptoms. Based on the literature in databases of China Journal Full-text Database (CNKI), Wanfang database, VIP network, China biomedical literature database (CBM) and PubMed, SPSS 20.0 and IBM SPSS Modeler 14.1 software were used to analyze the location of disease, symptoms, tongue and pulse, and syndrome type distribution through frequency statistics. In addition, association rule algorithm was used to explore the basic rules for underlying symptoms combinations of diabetic retinopathy. A total of 560 articles were retrieved, and a total of 240 articles met the inclusion and exclusion criteria. In these documents, 62 types of TCM syndromes were found, involving 9 820 cases, 201 symptoms, 29 tongue conditions, and 36 pulses conditions; the first two locations for the disease were liver and kidney. Then the association rule analysis of high frequency symptoms was used to dig out 15 groups of latent syndrome, and 3 underlying symptom combinations among high frequency symptoms, tongue conditions and pulse conditions. The results of the study showed that Qi and yin deficiency was most common for the diabetic retinopathy, and the location of the disease was closely related to liver and kidney. In addition, these high-frequency symptoms and tongue conditions, pulse conditions, and underlying symptom combinations can occur as main symptoms at diagnosis, providing reference for us to study the epidemiology of PRO scale of the disease. They can also increase the weight of these symptoms directly as the main symptoms, which can be also used as an alternative entry pool for TCM syndrome diagnostic scale, laying foundation for the construction and optimization of TCM symptom database of diabetic retinopathy.