In our life, music is one of the important tools of entertainment whose ingredients are musical instruments. For instance, acoustic drum is very important when a song is being sung. In the modern world, therefore, the style of the musical instruments is changing, yet retaining the same tune for instance an electronic drum. Based on MEMS 3D accelerometersensor data and machine learning, we have created “Virtual Musical Drums” in this work. Machine learning is integrating in all fields of AI for solving problems and the MEMS sensor is transforming a large ph This study has shown that, in the specified simulation, there is a detection accuracy of 91.42%, and in the real-time scenario with 20% window overlapping, there is an accuracy of 88. 20%. The simulated drum sound seemed unreal, even if the accuracy of the system detection was adequate. Therefore, we selected the "virtual musical drums sound files" in line with the acoustic drum sound pattern and length, and we adopted the "multiple hit detection within a fixed interval, sound intensity calibration, and sound tune parallel processing." Finally, but just as importantly, we completed the "Playing Virtual Musical Drums" exercise and were able to simulate playing an acoustic drum. Another aspect of MEMS sensors and machine learning has been illustrated by this work. It illustrates how data,. It demonstrates how data, sensor, and machine learning can be applied in a different way in offering musical entertainment with enhanced precision.