Forecasting stock prices is a challenging task due to the intricate dynamics of financial markets. Traditional models like ARIMA and conventional machine learning, such as decision trees, provide insights but struggle to capture the nuanced interactions affecting prices. In this study, we propose a novel approach by integrating news sentiment analysis and machine learning for advanced stock price prediction. We apply both ARIMA and decision trees to forecast stock prices, showcasing their proficiency in capturing temporal trends. Acknowledging the limitations of these methods, we delve into news sentiment analysis, extracting sentiments from company-related news headlines to tap into impactful external factors. Comparing sentiment-based machine learning with traditional techniques, our study highlights the superior predictive potential of sentiment analysis. The main contribution lies in emphasizing the significance of integrating external factors, such as news sentiment, for accurate stock price prediction. We demonstrate the promising performance of sentiment-based machine learning over traditional models, laying the foundation for hybrid techniques that merge traditional features and external sentiments. These hybrid approaches show great promise in providing more precise forecasts in the complex landscape of stock market dynamics.