Visually impaired people have many difficulties in their daily lives. In particular, they have difficulty signing the contract documents and application forms on their own at public institutions and financial institutions, but there are not enough administrative services to support document signing. Therefore, this study proposes a signature assistance system for visually impaired people using image analysis and voice guidance. The proposed system detects the signature field and the pen tip position, and instructs the user to place the pen tip over the signature field by voice guidance. The signature field is identified based on the positional relationship to the characters detected by YOLO. Some experimental verifications are provided to show the effectiveness of the proposed signature assistance system.