Ovarian cancer is a highly fatal gynecological malignancy with diverse subtypes—Clear Cell (CC), Endometrioid Carcinoma (EC), High-Grade Serous Carcinoma (HGSC), Low-Grade Serous Carcinoma (LGSC), and Mucinous Carcinoma (MC). Accurate subtype classification is crucial for effective treatment and prognosis but is hindered by the time-intensive, expertise-dependent process of manual histopathological examination. This study introduces a deep learning framework using a fine-tuned VGG16 model to automate subtype classification, trained on 24,965 histopathological images from The Cancer Genome Atlas (TCGA). The model achieved an accuracy of 72.7% on an independent test set, outperforming traditional methods. Grad-CAM visualization enhances interpretability, providing insights into the models decision-making. Our framework represents a significant step forward in ovarian cancer diagnostics, offering a reliable and interpretable AI solution to support personalized treatment strategies and improve patient outcomes.