Sentiment analysis of financial news headlines influences stock market values. We analyze this impact by collecting financial news and stock data over time. Using VADER for sentiment analysis, we establish correlations with stock prices sourced from Yahoo Finance and Kaggle, while headlines come from the Wall Street Journal, Washington Post, and Business-Standard. Text embeddings like BoW and TF-IDF are used, followed by machine learning (Naive Bayes, XGBoost) and deep learning models (CNN, LSTM). CNN (80.86%) and LSTM (84%) outperform SVM (50.3%), Random Forest (67.93%), and Naive Bayes (59.79%). BERT (90%) and RoBERTa (88%) achieve the highest accuracy.