Music evokes various emotions reflecting physiological and psychological changes within us. With the advancement in the area of music, along with effective computing, we can achieve the best music therapy solutions for various medical issues that are arising most commonly. Here, we develop an automated system for classifying the music genres with various algorithms like K-nearest neighbors, Support Vector Machines and Convolutional Neural Networks. Also, we derived audio features such that Mel-Frequency Cepstral Coefficients, zero crossing rate, and Spectral features from our suggested model. Our research solved genre classification problem with 94% accuracy and 0.94 F1 score using the convolutional neural network approach. The confusion matrix has been plotted and it is observed that there was a significant decrease in the loss (training-17%, testing-20%) with an increase in the number of epochs to 600.