OBJECTIVE: Temporomandibular joint osteoarthritis (TMJOA) has been modeled in different ways with a lack of uniformity. We aimed to investigate four TMJOA mouse models and assess histopathological changes in condyles, which could assist in the selection of animal models in further TMJOA-related studies. DESIGN: Four TMJOA mouse models were established, including unilateral hyperocclusion, discectomy, monosodium iodoacetate injection and aged model. Bilateral condyles were collected at different time points. The condylar alterations were analyzed by Micro-CT, Hematoxylin and eosin staining, Toluidine blue staining, Safranin O staining, Trap staining, immunofluorescence staining, immunohistochemistry staining and quantitative polymerase chain reaction. RESULTS: Radiographic and histopathological analysis indicated that all four methods could cause condylar degeneration successfully. Differences in morphologic and histologic changes were found among four models. The hyperocclusion model was time-dependent and the lesions got worse over time. Discectomy model presented obvious damage of cartilage and subchondral bone. Injected model showed severe inflammation and chondrocyte hypertrophy. The aged model was characterized by decreased of proteoglycan and osteolysis. CONCLUSIONS: The four methods had different characteristics and applicability. The harvest time affected the degree of cartilage degradation. Hyperocclusion was suited to explore the early-stage of TMJOA. Discectomy present advantages in investigating the long-term restoration of cartilage and subchondral bone. Monosodium iodoacetate-injection was appropriate for screening the agents for inflammatory relief. The aged model more naturally facilitated discovering underlying mechanisms in primary TMJOA. Unilateral modeling methods could initiate contralateral condylar alterations. The TMJOA models should be selected based on experimental requirements and applicability of each model.