With the rapid expansion of high-speed rail (HSR) network and the continuous improvement of passenger demand, how to scientifically evaluate and optimize the service quality of HSR operation has become a key problem to be solved urgently. This paper aims to build a set of methods for evaluating and optimizing the service quality of HSR operation based on intelligent technology, and put forward a multi-objective decision-making model to improve the service level in all directions. In this study, a multi-objective optimization model is constructed to improve passenger satisfaction and effectively control operating costs. The results show that the accuracy of BERT model in emotion analysis task is as high as 92.3%, which is significant compared with traditional methods. After optimization, the punctuality rate of trains increased from 89.5% to 93.2%, and the cleanliness score of carriages increased from 7.8 to 8.6, while the operating cost only increased by 4.7%. The research shows that the intelligent assessment and optimization method proposed in this paper can effectively find out the weak links of service quality and provide scientific improvement schemes, which provides an important reference for the decision-making of HSR operating enterprises.