The swift development of cyber threats, especially, sophisticated and distributed attacks on software-defined networks (SDNs) have necessitated smart and proactive threat-hunting solutions. This paper suggests a Hybrid Classical-Quantum Neural Network (C-QNN) model that combines classical machine learning algorithms with the near-term quantum computing capacity to improve predictive cyber threat detection. The proposed framework uses classical data mining techniques to extract features, reduce dimensionality and preprocess traffic pattern prior to passing optimized representations to variational quantum circuits to perform more sophisticated pattern recognition. The hybrid pipeline combines classical optimization techniques and quantum embedding strategies to overcome the drawbacks of existing quantum hardware and exploit quantum-enhanced learning capabilities. Empirical analysis shows better detection accuracy and predictive capabilities in detecting complex attack patterns in high dimensional network settings. The results emphasize the prospects of hybrid quantum-classical AI systems as a viable solution to deploy scalable, intelligent cybersecurity systems without depending on full fault-tolerant quantum processors. This study helps to establish the basis of embedding quantum-enhanced analytics into the next-generation proactive systems of cyber defense.