Alzheimer's disease (AD) is a progressive neurodegenerative condition that afflicts millions across the globe with cognitive impairment and memory loss. Early detection is important for effective intervention and enhanced patient care. This research examines a hybrid methodology that uses Convolutional Neural Networks (CNN) for feature extraction and Support Vector Machine (SVM) for classification to boost the accuracy of Alzheimer's identification. The dataset, obtained from the OASIS database, contains MRI scans and clinical measures. Following preprocessing and normalization, CNN was used to learn deep spatial features from brain images that were subsequently passed as input to the SVM classifier. The model had 98% accuracy with high precision, sensitivity, and F1-score compared to traditional machine learning models. The combination of SVM and CNN takes advantage of deep learning's feature extraction ability and the interpretability and efficiency of traditional classifiers. This hybrid scheme shows promise for more sensitive and accurate Alzheimer's diagnosis. Testing larger data sets and incorporating more clinical biomarkers should be addressed in future work.