Intrusion leads to violations of the security policies of a computer system. An intrusion detection system (IDS) is a software application that monitors network or system activities for pernicious activities. Many researchers propose the intrusion detection based on machine learning techniques or neural networks, but some of them didn't introduce high detection or decrease the time. The proposed framework is based on machine learning algorithms. These algorithms, discernibility classifier based k-nearest, J48 decision tree and Naïve Bayes rule, are used to discover any intrusion based on anomaly detection. The primary aim of this paper is to enhance the strength of the overall classification decision in better results than any other existent techniques. The performance metrics in our experimental are accuracy, error rate, sensitivity, specificity, and Precision. We notice during experimental results by using NSL-KDD data set, there are improvements in almost results by using the proposed framework.