This study presents a deep learning–based framework for the automated classification of retinal diseases using Optical Coherence Tomography (OCT) images. Convolutional neural network architectures, including ResNet50, Xception, and Inception V3, were developed and evaluated to distinguish between pathological and normal retinal conditions, such as Choroidal Neovascularization, Diabetic Macular Edema, and Drusen. The proposed models demonstrated high accuracy and strong generalization across benchmark OCT datasets. Incorporating preprocessing steps such as denoising significantly improved performance, particularly for the Xception and Inception V3 models. These findings highlight the potential of AI-driven analysis to support early diagnosis and clinical decision-making in ophthalmology.