This paper constructs a pyramid structure hierarchical model through fast non-dominated sorting and adaptation grading, realizes intra-layer competition and collaboration, and finds the optimal frontier. Combining entropy weight method and fuzzy comprehensive evaluation, the influence weights of external environment and individual ability are quantified, and a multi-objective optimization model of higher education talent cultivation path is established. The improved ant colony algorithm is introduced to enhance the global search capability of multi-objective optimization by using the state transfer rule and pheromone updating mechanism. Through the practice of higher education talent cultivation path optimization, as well as multi-algorithm comparison experiments, the optimization effectiveness of this paper’s method is verified. The results show that: using the method of this paper to determine the optimization of higher education talent cultivation path is divided into 3 stages, and the probability of achieving ability cultivation in each stage is 0.4, 0.3, 0.3, respectively. The method of this paper takes between 150-200ms to run in 5 test functions, and only needs a small number of iterations to get a stable fitness value. Using the method of this paper to continuously optimize the training path, the students’ employment rate, the rate of obtaining professional-related certificates, and the enterprise satisfaction rate are increased by 14%, 36%, and 11%, respectively.
