Alzheimer’s disease (AD), an irrevocable brain dis- ease, decreases thinking and memory power while the whole bit of mind size is pulled down, which at last reduces. It is also a neurodegenerative disorder and a very popular type of dementia in aged persons. There is a new case of Alzheimer’s disease being discovered globally every four seconds. Early detection and prediction of Alzheimer’s disease are extremely challenging. Timely identification of Alzheimer’s can be beneficial to get necessary care and even possibly prevent brain tissue damage This issue can be resolved by a machine learning system that has early disease prediction capabilities. This paper analyzed a number of previous research that used machine learning algorithms to diagnose Alzheimer’s disease over the previous three years. Comparisons are provided on the algorithms, assess- ment processes, and the obtained results. However, because key variables like feature selection and quantity impact the model’s performance and accuracy, the same algorithm’s accuracy may vary from dataset to dataset. An additional crucial finding in this review is that the ensemble models outperform regular models in terms of accuracy and performance. Future research can focus on merging numerous types of data sources, such as neuroimaging, genetic information, clinical reviews, and wearable device data. Integrating these diverse modalities can lead to more comprehensive and accurate predictive models.