In order to meet the requirements of vehicle load modeling and analysis for the digital transformation of highway bridges, the characteristics of vehicle load are analyzed and summarized. Based on the classification method of the number and complexity of random factors, vehicle load models are divided into single parameter vehicle load model, vehicle load spectrum, random traffic load model and vehicle load theoretical model. Based on this classification, the characteristics, modeling methods, development status, existing problems and application prospects of various vehicle load models are discussed. The findings demonstrate that existing vehicle load modeling methods mostly rely on sampling analysis of unknown stochastic processes, which are often inefficient and limited, and the resulting load models are not universal. In practical applications, appropriate modeling methods should be selected based on the characteristics of single factor vehicle load models, vehicle load spectra, and random traffic load models to meet application requirements. The stochastic traffic flow model is currently the most comprehensive modeling description of vehicle loads, and its basic research direction focuses on covering more load parameters, clarifying the correlation between various parameters, and more efficient modeling methods. The development of vehicle load theoretical model is not mature, which is the main direction of further research on vehicle load. With the development of society, the establishment of universal and automatic vehicle load model has become the basic trend of model development.