Currently, Alzheimer’s disease is one of the leading causes of dementia worldwide, and its early diagnosis remains a major challenge, especially in regions with limited resources. This study proposes a robust model based on deep learning techniques for the detection of Alzheimer’s disease using magnetic resonance imaging (MRI). The methodology was developed in four phases: data acquisition (using a set of more than 11,000 images categorized according to the degree of cognitive impairment), preprocessing (normalization and resizing), application of deep learning models (ViT, DeiT, Swin Transformer, EfficientNet, ConvNext, MobileViT, and PiT), and evaluation using metrics such as precision, accuracy, recall, and F1-score. After being refined, the DeiT model obtained the best results with 99.14% in all key metrics, demonstrating an almost perfect ability to correctly classify Alzheimer’s cases. The results show that optimized deep learning models have high potential to support the early diagnosis of this disease, facilitating more timely and effective medical interventions.