Objective To investigate the diagnostic value of multimodal fusion based on magnetic resonance imaging (MRI) combined with computed tomography (CT) three-dimensional (3D) reconstruction for structural knee joint injury, and to evaluate its application effects in clinical localization, lesion identification, and preoperative assessment, so as to provide technical support for accurate imaging assessment of complex knee lesions. Methods A total of 78 patients with knee joint injury admitted to Coast Guard Corps Hospital of Chinese People’s Armed Police Force from January 2023 to January 2025 were retrospectively analyzed. All patients underwent 64-slice CT and 1.5T MRI examination. Mimics software was used for three-dimensional reconstruction of CT images, and MRI was used for spatial registration and fusion. Using arthroscopic or intraoperative findings as the gold standard, the diagnostic efficacy of MRI alone, CT alone, and MRI+CT three-dimensional fusion in the identification of meniscus tears, cruciate ligament injuries, cartilage lesions and occult fractures was evaluated, and their performance in image quality, structural restoration and preoperative application were analyzed. Results The sensitivity, specificity and accuracy of MRI combined with CT three-dimensional reconstruction were 96.63%, 92.44% and 94.87%, respectively, which were higher than those of MRI (89.29%, 83.08%, 86.22%) and CT alone (81.65%, 79.36%, 80.64%)（all P<0.05). The Kappa value was 0.889, which was highly consistent. Receiver operating characteristic (ROC) curve analysis showed that the AUC of combined examination was the highest (0.933), and its diagnostic discrimination ability was the strongest (P<0.05). In 78 cases with successfully constructed 3D models, the image integrity score was (8.71±0.62), and the lesion localization error was (2.13±1.04) mm. Conclusion Multimodal fusion based on MRI combined with CT three-dimensional reconstruction has higher accuracy and consistency in the diagnosis of structural knee joint injury, and the fused image model shows significant advantages in clinical judgment and surgical planning. It is an effective means to achieve precision diagnosis and treatment and intelligent image assistance, which is worthy of clinical application in sports medicine and orthopedics.