Emotions of people can be different and are affected by internal and external factors. Many studies and research works have been undertaken concerning human emotion which in turn has resulted in a wide range of applications. Now, music playlists are mostly done automatically by genre or artists, but some music arrangements still need to be organized manually. This process also might take some time and not be liked by users which is some of its disadvantages. A brand new, innovative and advanced system we have introduced. most recently, mood-based music systems have incorporated deep learning methods. In this study, we come up with a new mood-based music system built on top of Inception and face recognition technologies. The program immediately analyzes an expression of the user to establish the emotion. Our tests prove that the proposed system is a very good tool for recognition of emotions and engaging the user with music that fits the current emotional state. These findings will showcase how deep learning methods such as Inception and facial recognition technologies can help to enhance the user experience in mood-based music systems. The performance of the proposed model was evaluated using different deep learning architectures such as MobileNetV2, VGG16, CNN, and ResNet152V2, achieving classification accuracies of 98.13%, 92.08%, 97.89%, and 96.67% respectively.