Background. Rheumatoid arthritis (RA) is an autoimmune disease characterized by multijoint swelling, pain, and destruction of the synovial joints. Treatments are available but new therapies are still required. One source of new therapies is natural products, including herbs used in traditional medicines. In China and neighbouring countries, natural products have been used throughout recorded history and are still in use for RA and its symptoms. This study used text-mining of a database of classical Chinese medical books to identify candidates for future clinical and experimental investigations of therapeutics for RA. Methods. The database Encyclopaedia of Traditional Chinese Medicine (Zhong Hua Yi Dian) includes the full texts of over 1,150 classical books. Eight traditional terms were searched. All citations were assessed for relevance to RA. Results and Conclusions. After removal of duplications, 3,174 citations were considered. After applying the exclusion and inclusion criteria, 548 citations of traditional formulas were included. These derived from 138 books written from 206 CE to 1948. These formulas included 5,018 ingredients (mean, 9 ingredients/formula) comprising 243 different natural products. When these text-mining results were compared to the 18 formulas recommended in a modern Chinese Medicine clinical practice guideline, 44% of the herbal formulas were the same. This suggests considerable continuity in the clinical application of these herbs between classical and modern Chinese medicine practice. Of the 15 herbs most frequently used as ingredients of the classical formulas, all have received research attention, and all have been reported to have anti-inflammatory effects. Two of these 15 herbs have already been developed into new anti-RA therapeutics—sinomenine from Sinomenium acutum (Thunb.) Rehd. & Wils and total glucosides of peony from Paeonia lactiflora Pall. Nevertheless, there remains considerable scope for further research. This text-mining approach was effective in identifying multiple natural product candidates for future research.