This study proposes a hybrid model combining Particle Swarm Optimization (PSO) and Long Short-Term Memory (LSTM) deep neural networks to forecast the stock prices of Apple, Amazon, Google, Microsoft, and Vinamilk. Currently, machine learning models such as LSTM, Gated Recurrent Unit (GRU), and Support Vector Regression (SVR) are widely applied in stock price prediction due to their ability to handle nonlinear data and complex market fluctuations. However, selecting suitable hyperparameters for these models remains a major challenge, directly impacting prediction accuracy. To address this issue, PSO is integrated to search and optimize key hyperparameters of the LSTM network, such as the number of hidden layers, learning rate, and batch size. Experimental results on stock price datasets demonstrate that the PSO-LSTM model outperforms traditional LSTM and other benchmark models regarding accuracy and stability. The proposed model not only enhances forecasting performance but also has the potential to become an effective decision-support tool for investors in the highly volatile stock market.