We present an analysis of optimal Convolution Neural Network (CNN) for fashion data classification by altering the layers of CNN in this paper. The suggested system employs three deep convolution layers, max pooling layers, and two fully connected layers, as well as dropout layers. While modified layers enhance the test accuracy of Adam optimizer when compared to start-of-art-models. The objective of this work is to address the multi class classification problem and to evaluate the performance of CNN’s Adam and RMSProp optimizer. The experiment was carried out using the Fashion-MNIST benchmark dataset. The suggested method has a test accuracy of 92.68%, compared to 91.86% in CNN using the softmax function and 92.22% in CNN utilizing batch normalization.