Breast cancer is one of the world’s most serious diseases that affect millions of women every year, and the number of people affected is increasing. The only practical way to lessen the impact of a disease is through early detection. Researchers have developed a variety of methods for identifying breast cancer, and using histopathology images as a tool has been quite successful. As an enhancement, this research develops a jelly electrophorus optimization-based 3D density connected deep Convolution Neural Networks (CNN) for the identification of breast cancer using histopathology images and input is collected and pre-processing is performed for improve the histopathology image’s properties. Then the feature extraction is performed through VGG-16, AlexNet, ResNet-101, statistical features. Finally, these concatenated features are fed forwarded to 3D density connected 3D deep Convolution Neural Networks (CNN) classifier it automatically detects breast cancer effectively with the help of jelly electrophorus optimization. The proposed jelly electrophorus optimization adjusts the classifier’s weights and bias. In terms of accuracy, sensitivity, and specificity, the proposed JEO-3D deep Convolution Neural Networks (CNN) approach attains values of 93.73%, 95.31%, and 94.14% for dataset 1 and 88.75%, 90.32%, and 89.15% for dataset 2, which is more successful.