The aim of the paper is to study and design a CAD tool for Alzheimer’s disease (AD) detection using linear SVM classifiers on MRI scans. Existing approaches often face challenges with high-dimensional feature spaces and limited data accessibility from repositories like ADNI. To overcome these, three SVM-based classifier proposals were explored. Results demonstrated promising accuracy, sensitivity, and specificity, particularly with cross-validation techniques. Notably, classifiers trained with image histograms showed superior performance in differentiating between AD, cognitively normal (CN), and mild cognitive impairment (MCI). Despite successes, initial approaches encountered difficulties, necessitating alternative methods. Challenges in data access and processing highlight the need for improved data management tools. Moving forward, refinements such as feature selection and testing with larger datasets could enhance classifier performance. Exploring deep neural networks as an extension may offer further insights into AD detection methodologies for early diagnosis and intervention.