Cybersecurity assurance in information n and communication systems has become a critical challenge due to the increasing complexity, scale, and persistence of modern cyber threats. Widely deployed security solutions, such as intrusion detection and prevention systems (IDS/IPS), primarily rely on reactive mechanisms and signature-based analysis, which limits their ability to detect novel and previously unseen attacks at early stages. This shortcoming highlights the need for proactive and intelligence-driven detection approaches. This paper proposes an intelligent proactive cyber threat detection model based on Cyber Threat Intelligence (CTI) and machine learning techniques. The proposed approach integrates heterogeneous CTI data sources with network and system security logs to extract behavior-oriented threat features. These features are processed through a multi-stage data preprocessing pipeline, including normalization, feature selection, and vector construction, enabling effective input for machine learning-based threat prediction. The developed algorithm estimates threat probabilities and computes dynamic risk scores to trigger early security alerts before full-scale attacks materialize. Experimental evaluation conducted in a simulated information and communication environment demonstrates that the proposed model outperforms conventional reactive detection approaches in terms of detection accuracy, response time, and false alert reduction. The results confirm the effectiveness of combining cyber threat intelligence with machine learning for proactive security analytics. The proposed solution is applicable to network security monitoring platforms, governmental and corporate information systems, and security operation centers, and it provides both practical value and a scalable foundation for further research in proactive cyber threat detection.