The stock market is characterized by high volatility and nonlinear patterns that make accurate price predictionachallenging task. This paper presents an empirical study on stock market prediction using machine learning, focusingonSupportVector Regression (SVR) for modeling and forecasting future price trends. Historical stock data, including highandlowpricemovements, are collected and preprocessed to remove inconsistencies and normalize input variables. The systemarchitectureintegrates feature extraction based on technical indicators and applies optimized SVR models to capture hiddendependencieswithin the data. Comparative analysis between the proposed real-time SVR model and a conventional SVMbaseline demonstratessignificant improvement in prediction accuracy and error reduction across evaluation metrics such as MSE, MAE, RMSE, andoverall trend accuracy. The experimental results indicate that the SVR model achieves superior generalization androbustnessincapturing complex stock dynamics, offering valuable insights for investors and analysts in data-driven financial decision-making.Keywords: Stock Market Prediction, Machine Learning, Support Vector Regression (SVR), Support Vector Machine(SVM),Financial Forecasting, Technical Indicators, Time Series Analysis, Price Trend Prediction, Data Preprocessing, ModelOptimization.