The emerging development of the internet everyone using web applications for their products or services. The large number of web applications created day by day. Since the demand is very high for web development, developers are creating an application not in a secure manner and hosting without the testing process. Web clients regularly store and oversee basic data that draws in cybercriminals who exploitation the web weaknesses for their benefits. Pernicious website pages are coming to pass undermining issue over the web on account of the reputation and their capacity to impact. Recognizing and examining them is exorbitant due to their characteristics and complexities. The complexities of assaults are expanding step by step in light of the fact that the assailants are utilizing mixed methodologies of different existing assaulting procedures. Using this opportunityattacker used their malicious script in their web application. Attacker, theft user’s data, or redirect to malicious websites. In this project, we are proposing detection methods, to prevent the users from approaching the malicious web application. Using a Machine learning algorithm, extract the feature of the web application that is URL features and static features of the network. From the trained model of date set, using the Random forest algorithm detect the malicious web application.