Millions of individuals worldwide suffer from Alzheimer’s disease (AD), a neurological illness that damages cognition and causes memory loss. Early and accurate diagnosis of Alzheimer’s disease is crucial for timely management and improved patient outcomes. This study investigates a unique approach for early AD identification using machine learning techniques on handwriting samples, an often-overlooked but potentially valuable source of diagnostic information. Handwriting analysis has shown promise in capturing subtle cognitive impairments that may precede clinical symptoms. The methodology involves developing a robust machine learning model that integrates advanced feature selection techniques to enhance efficiency and interpretability. To extract relevant characteristics from the handwriting data, we employ state-of-the-art feature selection methods, such as Recursive Feature Elimination (RFE) and LASSO regularization, to identify the most discriminative attributes. The CatBoost classifier is used to evaluate the AD diagnosis based on handwriting analysis. Several machine learning classifiers, including SVM, LR, NB, RF, DT, kNN, LDA, GLVQ, LGBM, XGBoost, and CNN, are assessed using standard evaluation metrics such as precision, accuracy, F1-Score, and recall.