Emotions have a strong impact on behavior. In face to face learning emotion can be accessed using non-verbal behavior, but in e- learning, it is challenging. We provide an audio as well facial data analysis engine to predict student’s behavior and help teachers provide assistance accordingly. System extracts features of linguistic as well as prosodic features. The system uses words and other features like pitch as these solely could not be as accurate as required, for example, if a student is happy or excited, his pitch is high and also if he is angry, his pitch is also high so we are using words, and on the basis of their combination we would extract the emotions of students. Also, the system tag students’ emotions for historical data. We also do emotions analysis from facial extraction/gesture analysis where data is collected through a camera attached to each workstation. We identify behavior of students during his study period to analyze their interest towards that a particular subject. Moreover, we identify after what time during studies a student loses interest in a particular subject and learning gets slower and sometimes stops. Our proposed solution consists of system attached with camera which is to record the video of students. Our objective of getting regular pictures of classrooms and record emotions of students while learning a subject. We label emotions of student using our application and record them.