Early and accurate prediction of Alzheimer’s disease (AD) is very important for improving patient care and proper management. This work presents a comparative performance evaluation of three different Machine Learning (ML) algorithms; Convolutional Neural Network (CNN), Random Forest (RF), and Logistic Regression (LR) for AD detection using structured clinical data. The dataset obtained from Kaggle comprises numerical features including age. Mini-Mental State Examination (MMSE) scores, Activity Daily Living (ADL), and other cognitive indicators. Python programming language was used as the tool and libraries like; pandas, numpy, sklearn, keras, tensorflow, statsmodel and others were used on jupyter notebook to implement the models. After preprocessing, and feature selection, the models were trained and evaluated using standard classification metrics. Results showed that the CNN model significantly outperformed the RF and LR models, achieving an overall accuracy of 93%, with precision, recall and F1-score consistently above 0.90. The RF model exhibited moderate performance, while the LR model underperformed with an accuracy of around 46%. This study demonstrates that CNNs can reliably detect AD with structured health data, offering a scalable, non-invasive diagnostic alternative, particularly in resource limited healthcare settings. Future work should explore integrating genetic, imaging and longitudinal data to further enhance predictive accuracy.