A person's facial expression is a great window into their emotional state. Around half of what we say to one another is conveyed nonverbally, while the other half is expressed vocally. One of the biggest challenges in computer science right now is learning to automatically recognise people's faces. There are several uses for facial expression recognition (FER), and they aren't restricted to gauging a person's mood or state of mind. Additionally, it is being used in a variety of industries, including criminal investigation, holographic as well as smart healthcare, security & education systems, robotics, and entertainment. Medical professionals are finding that facial expressions help patients with bipolar disease, whose emotions fluctuate often. In this study, a method for automatic face detection is developed by combining a pretrained VGG-16 convolutional neural network (FD-VGG-16 with CNN) with four convolutional layers and two hidden layers. The findings of this research employ a Cohn-Kanade dataset that includes extra pictures of male and female faces expressing different emotions, including happiness, anger, fear, disgust, contempt, neutrality, sadness, and surprise. Python is utilized as the implementation tool for this project, which uses datasets from the Kaggle repository. Preprocessing, feature extraction, as well as classification are all components of this study's FD-VGG-16 with CNN procedure. The accuracy of 98 percent of the suggested method is shown in both tabular and graphical forms.