The toll cancer takes on society becomes heavier and heavier, eventually becoming a major societal issue. As the top cause of mortality among all women, breast cancer is the deadliest malignancy which threatens women's lives. Acceleration, accuracy, and cost savings are all possible with computer-assisted breast cancer diagnosis. Ultrasound imaging is widely used in diagnosing breast cancer. It has contributed in the reduction in mortality. Each breast is viewed twice by radiologists, resulting in two pictures. Deep Learning algorithms have been the primary focus of current research. An Artificial Neural Network that has more than one layer is known as deep learning. Human cognition and brain anatomy are mimicked by the approaches’ implementations. Instead of relying on hand-crafted features, deep learning learns from the data itself. Raw data are sent into Deep Learning, which then learns features on its own before generating labelled classes from the inputs. The CNN model is used and the goal is to improve the flow of information inside the pre-trained network by transforming different matrices by correctly identifying it. The research highlights the potential practical utility of the proposed model, while downplaying the significance of early diagnosis of breast cancer by classification problem. The research objective is to increase the accuracy to measure the correctness of classification among the two classes benign and malignant of breast cancer. The accuracy achieved is 99%.