The foremost military and security issues of the twenty-first century have been and will persistently be influenced by the strategic dynamics of cyberspace and terrorism invalid input. Nevertheless, multiple conflicting definitions exist for both cyberspace and terrorism, and a globally accepted definition for numerous cyber-related behaviours (such as cyber-terrorism, cyber-warfare, and cyber-crime) is lacking. Cyber-terrorism is frequently characterized as the “intersection of terrorism and cyberspace,” resulting in extensive interpretation and ambiguity. Cybercriminals are evolving their methods over time to circumvent security measures. Traditional methods are inadequate for identifying zero-day attacks and advanced threats. Various ML algorithms have been created to identify cybercrimes and counteract cyber threats. This research aims to evaluate many prevalent machine-learning approaches to identify significant cyber risks in cyberspace. Five principal machine learning algorithms are mostly used: Random Forest, Naïve Bayes, Decision Tree, Logistic Regression and KNN. We have conducted a concise examination to assess the efficacy of these ML algorithms in Cyberattack detection and classification also used an ensemble technique adaboost utilizing commonly employed and benchmark datasets of the NSL-KDD dataset.