Stock price prediction is a fundamental and important task in the field of finance. This paper introduces the use of Graph Attention Networks (GAT), incorporating stock price movements and investor’s sentiment to improve stock price prediction performance. We compare GAT’s performance against different deep learning models, including Artificial Neural Networks (ANN), Recurrent Neural Networks (RNN), and Long Short-Term Memory (LSTM) networks. The results highlighted that GAT consistently outperforms the other models across multiple metrics, particularly when sentiment data are incorporated. This use of GAT and sentiment analysis improved the prediction accuracy, demonstrating its potential as a more effective method for predicting stock prices. These findings highlight the importance of integrating the additional sentiment information to enhance model performance and provide a promising direction for future research in financial forecasting.