The importance of bridge technical condition assessment is not only to ensure traffic safety but also to ensure the project’s sustainable development. Therefore, problems exist in the traditional highway bridge technical condition evaluation standard, such as fixed weight value, cumbersome calculation, and intense subjectivity. A total of 146 bridge inspection reports in Heilongjiang Province were collected in this paper. Using Pearson correlation analysis for bridge components and bridge age and length, the features with strong correlation are identified as the basis for modeling, and a machine learning model is introduced to evaluate the technical condition of the bridge. The application effects of the BP neural network, support vector machine (SVM), random forest (RF), and particle swarm optimization-support vector machine (PSO-SVM) in bridge evaluation and classification are compared and analyzed. The results show that the essential parameters penalty factor c and kernel function g in support vector machine optimized by particle swarm optimization overcome the shortcomings of the SVM model, improving the accuracy of the assessment. It can be used as an effective means to evaluate the technical condition of bridges and provide scientific decision-making reference for the maintenance of bridges.