This dissertation investigates the application of AI models for stock price prediction, a critical and emerging research area given the rapid advancements in AI across various fields. Traditional statistical models like ARIMA and GARCH are widely used for time series forecasting but often fail to capture the non-linear patterns in stock market data. This study explores the potential of these models, alongside a hybrid ARIMA-GARCH model, to enhance prediction accuracy. Using historical stock data from 2018 to 2023, the models are evaluated using four key metrics: RMSE, MAPE, MAE, and R2. The findings indicate that traditional models outperform the hybrid model in long-term forecasting, highlighting the need for further research into more effective hybrid approaches.