Neural style transfer (NST) is a technique used to stylize an image with the style of another image. In the past, this required the skills of an artist, but later machine learning was used to tackle this problem. Various hand-crafted methods were employed to extract features from the image, but these local, global, or combined features still produced subpar results. The idea of image analogies was then utilized to translate one image’s style to another by understanding their relationship. However, this method couldn’t learn the content-based association between images. With the success of convolutional neural networks (CNNs) in image-related tasks like detection, CNNs were used in NST. This study investigates style transfer using four different CNN models such as AlexNet, MobileNet, VGG-16, and VGG-19. Six pairs of images, referred to as content and style images, were evaluated with these pre-trained models, and their performance was evaluated based on content loss, style loss, runtime, and memory consumption. Results showed that VGG-19 outperforms other pre-trained CNN models in terms of quality, as its content loss is lower than the style loss.