Abstract: Focusing on the shortcomings of traditional models，a pavement performance evaluation model based on support vector machine (SVM) was proposed to comprehensively evaluate the performance of asphalt pavement．The method to determine the asphalt pavement performance training set and training labels was also proposed． At the same time，three optimization models，namely cross-validation (CV)，particle swarm optimization (PSO)，and genetic algorithm (GA)，were used to optimize the penalty parameter C and the kernel function parameter g which have effects on the accuracy of models，and their accuracy rate can reach to 99. 60%，96. 67%，94. 77% respec-tively． The results show that the best parameters obtained by cross-validation optimization have the highest accura-cy． Finally，taking the 23 maintenance sections of Guangyun Expressway as an example，the support vector ma-chine model and Technical Evaluation Standard for Highway Technology are respectively used to evaluate compre-hensive performance of the pavement． The results show that the model proposed in this paper is more suitable for practical application．

Key words: asphalt pavement, performance evaluation, machine learning, support vector machine