Keylogger, a tool intended to record every keystroke made on the machine and offers the attacker the ability to steal large amounts of sensitive information without the permission ofthe owner ofthe message. The primary objective of this project is to detect keylogger applications and prevent data loss and sensitive information leakage. This project aims to identifY the set of permissions and storage levels owned by each of the applications and hence differentiate applications with proper permissions and keylogger applications that can abuse permissions. The keyloggers are detected using Black-box technique. Black-box approach is based on behavioral characteristics which can be applied to all keyloggers and it does not rely on the structural characteristics of the keylogger. This project aims to develop detection system on mobile phones based on machine learning algorithm to detect keylogger applications.