One of the leading causes of dementia among older people is due to Alzheimer’s disease and the detection of cognitive impairment is essential for preventing the premature disease. Magnetic Resonance Imaging (MRI) plays a crucial role in studying the pathological condition of patients. Several models are developed for detecting the disease, but the traditional methods failed to prove their accurate of detection and the identification of congenital observations remained as a challenging concerm. In this research, a search and hunt optimization-based stacked deep convolutional neural network (SH-StNN) model is proposed for the effective prediction of the disease. The segmentation performance and the tuning of hyper-parameters are enhanced by the incorporation of the search and hunt optimization algorithm. The prediction model overcame the potential impacts caused by the loss of features, improper segmentation, and the prediction of minute abnormalities. The experimental analysis is done using the Alzheimer’s Neuro-imaging Initiative ADNI repository dataset and the modelattained a higher accuracyof 98%. The developed approach performs better than the conventional methods for classifying the diverse stages of the disease improving the therapeutic efficiency.