Abstract: To address issues such as the inability to meet consumer personalized demands, the insufficient analysis of user natural language expressions in traditional manufacturing, and the lack of accurate feedback interpretation by businesses, this paper introduces an enhanced GPT model based on Kansei engineering, referred to as "Kansei engineering GPT". Further, a distributed customized manufacturing solution based on it is proposed. The solution fine-tunes and trains the GPT language model while integrating a sentiment mining approach based on Kansei engineering to quantify the emotional attributes of Kansei vocabulary and generate products that meet user personalization. Additionally, digital prototypes are blockchain-uploaded to protect user intellectual property and sensitive information. To address challenges in distributed manufacturing environment such as the lack of transparent management mechanisms, significant equipment differences, and diverse user demands, we introduce a task allocation and pricing mechanism using multi-attribute reverse auctions and blockchain technology. Finally, a blockchain-based distributed 3D manufacturing management platform driven by Kansei engineering GPT is developed. By comparing it with other task allocation methods, the feasibility of the task allocation strategy proposed in the paper in a distributed manufacturing environment is verified. Results demonstrate that the distributed customized manufacturing solution driven by Kansei engineering GPT can effectively realize the user personalized needs in a distributed environment.