With the rapid evolution of digital technologies, there is an exponential increase in cybersecurity threats that creates a huge risk for organizations and individuals alike. Conventional cybersecurity measures tend to fall behind the sophistication of contemporary cyber-attacks. This paper suggests a prediction-based control strategy for cybersecurity threat detection in an IEEE 9-Bus system power grid that tries to counter such challenges. By using advanced algorithms and real-time analysis, the proposed system detects and prevents potential threats. The suggested system applies Artificial Intelligence and Machine Learning techniques like Random Forest, LSTM and Kalman Filtering to detect anomalies, that detect unusual patterns in data which could be an indication of a potential security risk. Through continuous learning from developing patterns of attacks, the suggested system tries to be adapt-able to novel threats. For such adaptation functions, the system suggests applying hybrid AI models utilizing supervised and unsupervised learning to detect known and unknown threats effectively.