Financial markets are characterized by increasing dynamism and complexity, posing significant challenges to traditional prediction methodologies, particularly in nonlinear and highly volatile regimes. To address these limitations, we present a novel multi-agent model leveraging recent advances in deep learning and quantum finance theory. Our approach integrates Quantum Price Level (QPL) theory, Transformer-based sentiment analysis, and established technical indicators (e.g., MA, RSI) within a cooperative agent framework to enhance predictive accuracy. The model constructs a comprehensive state representation by fusing diverse data sources, including historical price fluctuations, textual sentiment extracted from financial news, and quantum-derived threshold levels. Specialized agents are designed to process these distinct modalities and execute specific market-related tasks, coordinated to optimize overall portfolio performance. Empirical results demonstrate the superior predictive capability of our approach compared to conventional baselines, highlighting the synergistic benefits of integrating quantum-inspired features and sentiment analysis within a distributed multi-agent architecture.