With the increase of running speed, high-speed trains have higher requirement for aerodynamic shape. The pursuit of excellent performance and beautiful aerodynamic shape tends to be an important direction in the development of new high-speed trains. Based on the current aerodynamic study of high-speed trains, aerodynamic shape optimization can be divided into two categories: shape modification optimization based on flow mechanism and shape optimization based on optimization algorithms. After a brief review of the current domestic and foreign work on these two optimization approaches, this paper focuses on a series of aerodynamic shape optimization work done by the author's team in recent years. In terms of the modification optimization based on the flow field mechanism, this paper focuses on the research and development of the appearance of the two main models, namely "CRH" and "Fuxing" , to discuss the idea of the modification optimization. It mainly discusses the optimization design of air conditioning cover, pantograph installation platform, windshield and bogie apron, which have obvious influence on the train resistance, and introduces the improvement of aerodynamic performance of these components compared with the previous models. In the shape optimization method based on an optimization algorithm, according to the aerodynamic shape optimization process, taking the streamline of high-speed train head as the main optimization object, we introduce our series of work from three aspects: high-speed train parameterization method, surrogate model development and the improvement of the optimization algorithm. Among them, high-speed train parameterization methods mainly include local function method, modified vehicle modeling function method, and category/shape function method; in the development of the alternative model, the optimization alternative model and Kriging model based on cross-validation are introduced; on the improvement of the optimization algorithm, the improved non-inferior classification multi-objective particle swarm optimization algorithm and continuous domain chaotic ant colony algorithm are introduced. Based on the improvement of the above three aspects, the application cases of the developed aerodynamic shape optimization strategy in typical engineering are further introduced.