Alzheimer's disease (AD) is a complicated, irreversible, incurable neurological illness that has a positive global impact on human existence. It was the sixth killer in the United States, and no immunizations were available. According to research, this fatal condition is incurable. The progression of the disease might be delayed by improving the patient's quality of life, consequently enhancing the patient's cognitive abilities. So we proposed EAD-HNN (Enhanced Alzheimer's disease Detection using hybrid neural networks) approach typically, an MR image of the brain is used to diagnose Alzheimer's. In this proposed system Noise removal using MLP with Histogram equalization, the segmentation has done with Edge based with Robert operator, the training has carried with CNN and RESNet50. The classification has done with CNN Algorithm. This enables the use of picture recognition techniques in many ways to promote and improve diagnosis. Automatic identification of any illness sample saves physicians time and improves their accuracy. This article, which discusses the functional criteria for the clinical diagnosis of Alzheimer's, proposes a weighted combination of positive and negative samples and a technique for learning a limited number of pieces to enhance the data set system. It builds a Deep learning model that gives enhanced image details and improves the model's generalization capabilities.