Breast cancer is a life-threatening disease that mostly affects women. It occurs in the form of breast cancer and early diagnosis is important for effective treatment. Factors may include age, gender, family history and genetics. The aim of this project is to detect normal, malignant and benign images. Data is collected from the Kaggle repository. This article uses the “Breast Ultrasound Imaging Dataset (BISI)” database. The file contains a total of 780 images divided into three groups: normal, benign and malignant. The data is divided into 70% for training and 30% for validation. This study uses the power of machine learning (ML) to improve cancer diagnosis. The main goal is to develop a cognitive model that can predict whether medical images contain signs of cancer. This method involves first processing the image using a Wiener filter to reduce noise, then feature removal using a gray-level joint computation matrix (GLCM) to capture the details of the subject. The far-reaching impact of this program is its ability to improve patient outcomes, providing hope for the trial period. Three classifications were used to classify the samples, the accuracy of the decision tree reached 97.6%, the accuracy of the random forest reached 99.4%, and the accuracy of the SVM reached 98.5%.