Generative AI (GenAI) has recently sparked significant advancements in artificial intelligence, drawing many researchers to study areas like dialogue, text, image, and video generation. This paper proposes an automatic music Creation system built on the Transformer model in deep learning, which can create music based on user-provided melody fragments and predefined music style parameters. The system consists of two primary components: a user configuration interface and a music generation server. Users can input melody snippets and adjust parameters such as stability, variability, and length through the interface. The music generation server then leverages these parameters and the deep learning model to produce music that matches the desired style and mood. Furthermore, this paper uses the Dynamic Programming algorithm to evaluate the correlation between the generated melody and the user’s input notes. Initial user feedback has been highly positive, indicating that the system effectively supports music creators in their work, enabling them to generate music more efficiently.