Music is a perfect thing to lighten one up and enjoy the time after a busy work week or a weekend. Not only as a medium for dancing it can be used as a medium for syncing with your feelings and emotions also. But people cannot just checkout the desired song, when they will need it. Although they are some kind of songs they would love to listen to but they need help with finding those songs. The objective of this project is to make people happier by providing them their much-loved genre of music with just a few clicks, by building an efficient music genre classifier which can classify music into different genres. In MIR, genre classification is situated in the background. This project is developed to train and construct several machine learning models using them to automatically segregate music based on various genres. The project utilizes three distinct algorithms: KNN, SVM, and CNN are the examples of the machine learning techniques. To this end, we have opted for the extensively employed GTZAN dataset that boasts of 1000 audio samples (100 samples representing 10 genres). Additionally, we utilize a dataset comprising various audio features. The system is built on the many audio properties of the song samples through waveforms, and mel-frequency cepstral coefficients (MFCCs) among others. Next, these characteristics are provided as an entrance for all models. The present project which includes prudent experimentation and analysis seeks to provide tacit knowledge about the relative prevalence of different machine learning approaches for music genre recognition, this is expected to be a step in the direction of making new generation of automated music classification systems.