Movies often come with genre classification and this information can often help consumers determine whether to watch a movie. However, movie genre classification is not necessarily conclusive and rigid, as tagging of movie genre is still a manual process that involves a collection of users’ suggestions. This project outlines the approaches for genre predictions of movies based on their metadata and posters. The traditional machine learning models were employed in addition to Convolutional Neural Networks. Movie metadata and posters were obtained from TMDB and IMDB databases.