The organic combination of telemedicine and technology classification provides new ideas for the development of intelligent diagnosis of ophthalmic diseases. In this paper, Gray Wolf optimization algorithm is used for ophthalmic image feature selection, and a M2LC-Net model is proposed to solve the multimodal and multidisease long-tailed data classification, and Grad-CAM visualization module is introduced. A performance comparison is made with the unimodal model to verify the superiority of multimodality. The application effect analysis with the existing bimodal model shows that the AUC of the M2LC-Net model is improved by about 14.03% over the ResNet50 model and by about 9.00% over the DenseNet121 model. The results show that the M2LC-Net model is superior in the classification of ophthalmic diseases, and it can provide better technical support for building an ophthalmic care service model based on telemedicine technology.