A novel approach, known as MS-SSA-LSTM, has emerged as a highly effective method for accurately forecasting stock prices. This technique leverages the integration of diverse data streams, incorporating sentiment analysis, swarm intelligence techniques, and deep learning principles. The approach involves scanning forum entries to establish a sentiment index, optimizing the hyper-parameters of LSTM using the Sparrow Search Algorithm, and using LSTM to forecast stock prices by combining sentiment data with fundamental trading data. Improved stock forecast accuracy would enable investors to make safer decisions and profit from the financial market more. This is the main objective. The sentiment index enhances the predictive performance of the model, and the application of SSA aids in modifying the hyper-parameters of the LSTM and offers objective justifications for the model’s predictions. Additionally, we incorporate an ensemble technique, which combines projections from multiple independent models, to enhance the precision and reliability of the final prediction. We may look at more ensemble ways to boost performance even further, including Voting Classifier and LSTM + GRU for stock sentiment (which yielded 100% accuracy) and Voting Regression for stock price (which yielded 99% R2 Score). We enhanced the sentiment analysis by adding several emotional cues, such as anger, fear, grief, and disgust. Flask framework is used in the front end development process for user authentication and testing.