Alzheimer’s disease impairs memory, cognition, and organization. This causes dementia most often and worsens with time. Even though there is no cure, early detection and intervention are essential for symptom management and disease progression. The disorder may impair mobility, swallowing, behavior, speech, and communication. Healthcare providers and organizations describe Alzheimer’s disease phases differently. Dementia severity characterizes preclinical Alzheimer’s disease, mild cognitive impairment, and mild, moderate, and severe dementia. Prognosis and illness monitoring require precise classification. Automated clinical trial algorithms and machine learning for clinical and biomarker data analysis have enhanced sickness tracking and differential diagnosis. Many researches have diagnosed Alzheimer’s disease using MRI scans and machine learning algorithms like support vector machine, neural networks, and others. This study employs machine learning algorithms to predict Alzheimer’s disease using MRI scans and psychological factors like age, frequency of visits, gender, socioeconomic status, education, and others. First, Pearson correlation and T-test assessed psychological factors’ influence on Alzheimer’s detection. Alzheimer’s disease detection and classification models were evaluated using ensemble learning with voting rules. A soft voting ensemble learning system uses random forest, k-nearest neighbors, decision tree, and extreme gradient boosting to detect Alzheimer’s illness. The system outperformed with accuracy of 95.7% individual models, with random forest for non-demented levels and xgboost for severe ones. The study also showed that age, gender, education, socioeconomic status, and mini-mental state examination affect Alzheimer’s severity. Though senior folks are more likely to get Alzheimer’s, educated people can take measures. Group quality improves with higher mini-mental state examination and clinical dementia rating scores.