Alzhemier’s Disease (AD) is a chronic, progressive, and neurodegenerative brain disorder that damages the thinking skills, cognitive function. Conventional methods does not provide the accurate classification of AD. This research proposed the DenseNet 121 with Global Attention Module (GAM) for classifying the AD by using MRI images. The Non-Local Mean (NLM) filter is employ to pre-processing the MRI image to denoising and enhance the image quality. Then, the Adaptive Mean Thresholding (AMT) is provide for segmenting the relevant brain regions in image. The Grey Level Co-Occurrence Matrix approach (GLCM) offers to quantify textural information from the MRI images. The dataset named alzheimer's disease multiclass images dataset is used which is available in kaggle. Finally, Attention DenseNet121 is proposed that provides the high-precise classification of AD. Using Python software, the proposed Attention DenseNet121 achieved a higher uniform value of 99.58% across the accuracy, precision, recall and F1-Score which is highly efficient compared to the existing method.