Alzheimer’s disease (AD) is a neuro disorder that progressively deteriorates memory and thinking skills. There are diagnostic procedures that are in place, but they do not work properly. Deep learning (DL) has a lot of promise in automatically diagnosing Alzheimer’s disease through neuroimaging data. In this work I propose a novel approach for multi-class Alzheimer’s disease prediction using transfer learning with VGG-19 model and three-dimensional (3D) magnetic resonance imaging (MRI) scans. A pre-trained convolutional neural network (CNN) based architecture, VGG-19, was specially trained to classify Alzheimer’s disease patients into control, moderate cognitive impairment (MCI), and other brain disorders within 3D MRI scans. The proposed method was able to learn from sophisticated MRI images and improve the performance on medical imaging tasks, thus eliminating the dependency on a large, labeled dataset. Experiments have shown that the model is effective in classifying AD stages and achieves the best accuracy compared to the rest of the models as well as computational efficiency. Transfer learning is predominantly used here to enhance the early diagnosis of Alzheimer’s disease using clinicians in patient management as well as decision support systems.