The increasing volume of malicious content in social networks requires automating methods to detect and eliminate malicious content the URLs. This shows a supervised machine learning classification model that has been built to detect malicious content in online social networks. Multisource features are used to detect social network posts that contain malicious Uniform Resource Locators (URL's). These URLs could direct users to websites that contain malicious content, drive-by download attacks, phishing, spam, and scams and some other problems. For, the data collected from such URL's, the Twitter streaming application programming interface (API) was used.