With the evolution of mobile technology and the dynamic expansion of online information, the distribution of fake news has become an important social challenge. Several studies are addressing the need to develop automated systems that can detect and track fake news on online social networks. The work in this paper proposes an efficient model for detecting fake news on online social networks by utilizing deep learning techniques. Given the rapid distribution of fake news on platforms such as Instagram, X, and Facebook, there is a demand for effective detection methods that can determine and moderate the impact of fake news. Our approach merges Natural Language Processing (NLP) for analysing text with social network analysis to integrate user interactions and network structure. By using advanced deep learning models. In this work, we aim to improve the accuracy of fake news detection using Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Transformers (BERT). We will assess our proposed model using a comprehensive dataset and evaluate its performance against existing methods.