This study presents a novel hybrid modeling approach that combines a parametric stochastic volatility model, referred to as the Optimal Control (OC) model, with Artificial Intelligence (AI) techniques to produce high-precision forecasts of the log returns (or prices) of risky financial assets. The OC model, inspired by the classical Heston model, is first calibrated on real market data to generate initial forecasts, which are then refined using analytical AI methods. This hybrid approach yields improved results, demonstrating its effectiveness in enhancing predictive accuracy. Numerical experiments based on daily S\&P500 index data from the New York Stock Exchange show that the hybrid models consistently outperform the standalone OC model. The AI-driven corrections, in particular, effectively capture abrupt market changes, delivering robust performance during both stable periods and turbulent market conditions such as the COVID-19 pandemic. The numerical results confirm the potential of hybrid modeling for financial forecasting and highlight its applicability across various markets, including commodities and foreign exchange