Ovarian cancer is one of the primary causes of death in women from malignancies. Despite being one of the most common forms of cancer in women, it usually goes unnoticed until it reaches the advanced stage because current tools in the detection of the disease at an earlier stage are in short supply. Hence, this study aims to fill the current gap in the development of accurate and reliable methods of diagnosis to improve the rate of early detection and outcomes. A total of 3500 MRI images were collected from hospitals and online sources representing cancerous and non-cancerous cases at various stages. Preprocess the images and train a set of deep learning models like VGG16, VGG19, ResNet50, and Inception V3. Their performance was graded using the measurement metrics of accuracy, precision, recall, and F1-score. Among all the models, Inception V3 performed the best at 97.87%, while for ResNet50, it reached 94.5%, for VGG19 at 92.3%, and lastly, for VGG16 at 90.23%. The research study shows that deep learning models can also be used to enhance the diagnostic accuracy of ovarian cancer. Inception V3 can be said to be the best model that mainly applies to real-time diagnostics. Such CNN-based models, as this study indicates, can be very helpful in clinical practice regarding effective early detection to help radiologists conduct an accurate diagnosis, which will eventually improve the patient’s outcome. It researches and readdresses the transformative potential of AI in medical imaging.