Aerospace systems are enormous in numbers and diverse, from small to complex subsystems. They range from a component to a complete air system as well as space systems, etc. A complete and universal machine learning (ML) framework for aerospace systems is still unavailable to this date. This paper proposes and implements an ML-based Aerospace System (MAS) framework to understand and model data of the targeted aerospace systems. Any of the aerospace systems can be modeled using the MAS framework. For easy presentation and convenience, three chosen aerospace systems are considered to be modeled using the MAS framework. The results show that the MAS framework has negligible error, and high accuracy. The proposed MAS framework is equally applicable to other remaining aerospace systems minimum modification. The MAS framework can be thought of as a universal artificial intelligence tool to understand, model, and predict the real-time behavior of aerospace systems.