In this paper, optical coherence tomography (OCT) and surface-enhanced Raman spectroscopy (SERS) were used to characterize normal knee joint (NKJ) tissue and knee osteoarthritis (KOA) tissue ex vivo. OCT images show that there is a clear hierarchical structure in NKJ tissue, including surface layer, transitional layer, radiation layer and cartilage matrix calcification layer tissue structure, while the hierarchical structure of KOA tissue is not clear and unevenly distributed, and the pathological tissues at different stages also show significant differences. SERS shows that NKJ tissue and mild osteoarthritic knee cartilage (MiKOA) tissue have strong characteristic Raman peaks at 964, 1073 (1086), 1271, 1305, 1442, 1660 and 1763$$cm${}^{?1}$. Compared with the Raman spectrum of NKJ tissue, the Raman characteristic peaks of MiKOA tissue have some shifts, moving from 1073$$cm${}^{?1}$ to 1086$$cm${}^{?1}$ and from 1542$$cm${}^{?1}$ to 1442$$cm${}^{?1}$. There is a characteristic Raman peak of 1271$$cm${}^{?1}$ in MiKOA tissue, but not in NKJ tissue. Compared with NKJ tissue, severely degenerated cartilage (SdKOA) tissues show some new SERS peaks at 1008, 1245, 1285, 1311 and 1321$$cm${}^{?1}$, which are not seen in SERS spectra of NKJ tissue. Principal component analysis (PCA) was used to analyze the Raman spectra of 1245–1345$$cm${}^{?1}$ region. The results show that PCA can distinguish NKJ, MiKOA and SdKOA tissues and the accuracy is about 90%. These results indicate that OCT can clearly distinguish NKJ, MiKOA, moderate osteoarthritic knee cartilage (MoKOA) and SdKOA tissue, while SERS can provide further judgment basis. The results also prove that the contents of protein and polysaccharide in knee tissue are changed during the pathological process of knee tissue, which is the cause of pain caused by poor friction in knee joint during movement.In this paper, optical coherence tomography (OCT) and surface-enhanced Raman spectroscopy (SERS) were used to characterize normal knee joint (NKJ) tissue and knee osteoarthritis (KOA) tissue ex vivo. OCT images show that there is a clear hierarchical structure in NKJ tissue, including surface layer, transitional layer, radiation layer and cartilage matrix calcification layer tissue structure, while the hierarchical structure of KOA tissue is not clear and unevenly distributed, and the pathological tissues at different stages also show significant differences. SERS shows that NKJ tissue and mild osteoarthritic knee cartilage (MiKOA) tissue have strong characteristic Raman peaks at 964, 1073 (1086), 1271, 1305, 1442, 1660 and 1763$$cm${}^{?1}$. Compared with the Raman spectrum of NKJ tissue, the Raman characteristic peaks of MiKOA tissue have some shifts, moving from 1073$$cm${}^{?1}$ to 1086$$cm${}^{?1}$ and from 1542$$cm${}^{?1}$ to 1442$$cm${}^{?1}$. There is a characteristic Raman peak of 1271$$cm${}^{?1}$ in MiKOA tissue, but not in NKJ tissue. Compared with NKJ tissue, severely degenerated cartilage (SdKOA) tissues show some new SERS peaks at 1008, 1245, 1285, 1311 and 1321$$cm${}^{?1}$, which are not seen in SERS spectra of NKJ tissue. Principal component analysis (PCA) was used to analyze the Raman spectra of 1245–1345$$cm${}^{?1}$ region. The results show that PCA can distinguish NKJ, MiKOA and SdKOA tissues and the accuracy is about 90%. These results indicate that OCT can clearly distinguish NKJ, MiKOA, moderate osteoarthritic knee cartilage (MoKOA) and SdKOA tissue, while SERS can provide further judgment basis. The results also prove that the contents of protein and polysaccharide in knee tissue are changed during the pathological process of knee tissue, which is the cause of pain caused by poor friction in knee joint during movement.