In the constantly evolving cybersecurity landscape, effectively categorizing cyber-threat actors based on attack patterns is critical for establishing effective defense systems. Cyberattack detection and categorization are crucial for effective threat attribution. This study proposes a novel methodology for categorizing attack patterns based on distinct attack attributes to improve cyber-threat actor (CTA) attribution. The system uses advanced machine learning techniques to analyze a large dataset of cyber-attacks, detecting unique patterns and attributes associated with different threat actors. This study employs a dataset containing a variety of attack features. It aims to categorize cyber-attacks like ransomware, phishing, spoofing, dictionary attacks, and man-in-the-middle attacks (MITM), among others from the dataset. This study aims to identify trends in attack vectors, allowing for a more exact categorization of cyber-threat actors based on attack patterns. Experimental results show that the framework accurately categorizes threats and has the potential to improve response times and precision in cyber threat attribution.