Age-related Macular Degeneration (AMD) is one of the most causes for elders’ vision loss, early screening and treatment are the most efficient way to reduce the rate of blindness. AI-based methods based on ophthalmic images play a gat potential for AMD diagnosis. However, the difficulty of computing device obtaining, multiple evidence of image sources, time-wasting, and low level of explanation are challenges for AI models applicated in clinics. Thus, this study proposed a fusion learning method for AMD detection. Three steps are involved, which are image feature extraction, feature matrix fusion, and MLP-based AMD classification. Unsupervised (Hierarchical Clustering, SVM, and ResNet-K Means), supervised (VGG-16 and ResNet) methods and the proposed method are compared based on Optical Coherence Tomography (OCT), Fundus Autofluorescence (FAF), regular fundus photography (RCFP) and Ultra-Widefield Fundus (UWF), respectively and comprehensively. Findings show that the proposed method presents a high performance for integrated ophthalmic image diagnosis, it is timesaving (0.09s per image) with high precision (0.95), sensitivity (0.93), specificity (0.92) and AUC (0.94). Thus, this study concluded that the proposed method is a solution to AMD automatic quick detecting based on multiple data sources. A real-world UWF database is involved from Shenzhen Aier Hospital. Practical and theoretical contributions are delivered. A reference value for medical diagnosis based on multiple digital images is contributed.