Facial beauty prediction (FBP) is a rising topic in the area of evaluating attractiveness, with the goal of making assessments which are more human-like. Since FBP is a regression problem, there are many data driven ways to determine the relationships among face traits and beauty ratings. Deep learning approaches have shown their incredible potential for feature representations and analysis of the data. Convolutional Neural Network (CNN) has demonstrated extremely outstanding performance in facial identification and understandings, and have been proven to be an effective way for exploring facial features. Recently, a lot of well-designed networks with efficient topologies have been researched for improved representation of ability. However, these models focus more on the efficient block but do not create an efficient data transmission route, resulting in a feature representation capacity which is sub-optimal. In this paper, we have taken the Alex Net model into consideration which is a deep convolutional network which has been created to consider a large range of colorful images (224x224x3). It has a range of 62 million parameters which are trainable. For better performance, an intricate network architecture for the FBP issue is given in this research.