<p>Optimizing the selection of concrete proportion is the key to ensuring the strength, durability and cost-effectiveness of construction projects. However, the traditional mixed proportion method is costly, time-consuming and highly inconsistent. Therefore, this study aims to explore a new method to optimize the existing specific mixing ratio. In this study, support vector machine regression algorithm is used to establish a function that maps the existing composition data to the strength of concrete, and accurately predicts the performance of concrete. Then the predicted performance index is passed to the genetic algorithm. The experimental results show that the fitness value of HPO–SVM is 0.021, which is significantly lower than that of SSA–SVM and PSO–SVM, indicating the superiority of HPO in optimizing SVM. In addition, the average error value of HPO–SVM in specific proportion prediction is 1.226, which verifies the stability and accuracy of the prediction. The results show that the enhanced genetic algorithm has obvious advantages in improving convergence speed and solution quality. The concrete produced by the optimized mixture ratio has stronger compressive strength and bending strength. This study provides not only new methodological achievements, but also has important application value and popularization prospects in engineering practice and education, which is helpful to realize the goal of sustainable development of the construction industry and conforms to the development trend of green environmental protection.</p>