Stock price prediction in emerging markets is challenging due to volatility, inefficiencies, and complex macroeconomic interactions. Traditional econometric models struggle to capture nonlinear dynamics, limiting their predictive power. This study proposes a hybrid AI-driven forecasting framework integrating Fuzzy Logic, Particle Swarm Optimization (PSO), and Explainable AI (XAI) with Machine Learning (ML) to enhance both accuracy and interpretability. Using data from 22 publicly traded Vietnamese real estate firms (2013–2024), the model combines financial indicators (e.g., book value, earnings per share) with macroeconomic variables (e.g., CPI, oil prices, Federal Reserve rates). The hybrid approach outperforms traditional models, achieving an R² of 96% compared to prior benchmarks (R² ≤ 82%). XAI techniques, particularly SHAP, reveal key stock price drivers such as firm size and interest rate fluctuations, offering deeper insights for investors and policymakers. This study extends the Fama-French and Arbitrage Pricing Theory (APT) frameworks, integrating AI-based feature interpretation for improved decision-making. The findings have significant implications for portfolio management, risk assessment, and regulatory oversight, equipping stakeholders with a robust tool to navigate market uncertainties.