Social media provides a vibrant environment for sharing content and communicating online, allowing users to express themselves freely. Text is one of the most important and widely used types of content on these platforms, and classifying text comments is important for understanding individuals’ thoughts and trends. The study aims to develop an effective model for analyzing and classifying comments on social media networks into negative, positive, and neutral comments using natural language processing techniques, deep learning models, and machine learning, with word embedding. Convolutional Neural Networks (CNN) are used to extract information from texts and apply the Support Vector Machine (SVM) model to classification accuracy. Twitter was chosen due to its importance nowadays, as it is considered the most influential and interactive platform on social media. The results showed a high accuracy of up to 98% for this hybrid model compared to the traditional machine learning model on the Twitter datasets available on Kaggle.