It is difficult to forecast the prices of stocks because financial markets are very complex, nondeterministic, and changeable. However, many researchers have explored different ways to address this challenge, and machine learning techniques have shown strong potential in this area. In the current paper, we investigate the use of Long Short-Term Memory (LSTM) networks to predict short-term movements in stock prices based on an analysis of historical data and technical indicators. LSTMs are particularly suited to this type of problem because they can learn patterns and relationships that unfold over time, which is crucial for understanding market trends. We proposed a predicting model and then conducted experiments to validate it. The performance of the model was assessed for a range of metrics and compared with results from other machine learning methods and conventional investment strategies. Our findings were encouraging, with the LSTM model able to achieve an average accuracy of approximately 84.9% in making predictions about whether the price of a stock would increase in the near future. With the stock market intrinsically being unpredictable, the results obtained show that deep learning models, particularly LSTMs, can learn meaningful patterns in stock data and are definitely helpful decision-making and investment research tools.