“The cure for the pain is in the pain” is one of the most well-known and significant quotations. Kidney stone disease is a prevalent urological condition affecting millions globally, necessitating effective diagnostic methods. Manual segmentation and classification in MR images are challenging due to high data volumes and blurred boundaries. This research introduces a method to enhance accuracy and reduce diagnosis time by automatically classifying tissues into normal, benign, and malignant categories. The proposed approach involves pre-processing images to remove noise, normalize intensity, and enhance key features. The Discrete Wavelet Transform (DWT), which highlights significant details and reduces the size of the image, is used to extract important things. Additionally, the Gray level Co-occurrence matrix [GLCM] analyzes texture to extract features such as contrast, correlation, and energy. A Convolutional neural network [CNN] is then employed for classification. This case focuses on developing an efficient CNN model for detecting kidney stones in medical images, thus automating the diagnostic process for faster and more accurate results. The research covers dataset creation, CNN design, training, and evaluation, with thorough validation to ensure effectiveness and reliability. By enabling early detection and treatment, the proposed system aims to improve patient outcomes, reduce diagnosis time, and optimize resource utilization in healthcare settings.