Alzheimer’s disease (AD) is an ageing-related disease that causes the brain to deteriorate over time. It begins mild, but over time, it develops increasingly more severe. AD causes damage to brain cells as well as the death of those cells. Magnetic Resonance Imaging (MRI) is crucial to many clinical applications, including procedures for early diagnosis of different medical conditions and investigations into the internal workings of the human brain. The Deep Learning (DL) technique has significantly classified and detected based on MRI scans and it can obtain an accuracy that surpasses human performance. MRI is an effective diagnostic technique for correctly identifying brain diseases. However, one of the significant issues with most MRI scans involves low contrast, noise, and darkness, as well as the difficulty of obtaining massive datasets owing to the patient’s privacy. As a result, the medical image is imbalanced, which leads to overfitting and poor generalization. Address these issues by utilizing the enhancement image techniques and augmentation algorithms. In this paper, we applied the Contrast Limited Adaptive Histogram Equalization (CLAHE) with the traditional augmentation technique to increase the image quality and diagnostic ability. We proposed a custom Convolutional Neural Network (CNN) to classify four AD classes with a high detection accuracy of 99%, the effect on classification performance was observed by using CLAHE. The outcomes demonstrate that the medical imaging enhancing technique is adaptable and practical. This research presented a paradigm for the diagnosis of AD at an early stage while causing no brain injury, considering the potential for use in other domains.