The invention discloses a rap music automatic generation method based on voice input. The method comprises the following steps that S1. a user inputs voice information; S2. the voice information inputted by the user is converted into text information; S3. word segmentation is performed on the text information so that multiple phrases are obtained; S4. candidate Riff is selected from a material library through screening according to the phrases obtained in the step S3; S5. segmentation is performed on the original voice information according to the phrases obtained in the step S3 and a voice endpoint detection technology so that segmented voice is obtained; S6. effectors are added in the segmented voice and the candidate Riff; and S7. rap music is outputted. The general public are enabled to participate in music production, interaction and other professional activities to create their own music through assistance of machine learning and other technologies, and the user only requires to input the voice information in the process so that the corresponding rap music can be automatically generated.