In order to explore the potential and challenges of the application of driving simulators in the study of driver behavior, this study adopts a systematic literature review and bibliometrics approach, taking 1110 documents from Web of Science core database and 240 documents from CNKI core database as samples, and using CiteSpace for data visualization and analysis. The study summarizes three major application scenarios of driving simulators: intelligent mining of driver behavioral features, construction and validation of driving behavior models, and analysis of driving safety influencing factors. In terms of intelligent mining of driver behavioral characteristics, the study focuses on the application of driving simulators in the study of driver interaction with assisted driving functions and automatic driving functions; in terms of the study of the influence of driving safety, the analysis is carried out from the three dimensions of the driver"s own condition, the driving environment, and the driving training; and in terms of the construction and validation of driving behavioral models, the representative models and the related experimental designs are discussed in depth. In the construction and validation of driving behavior models, representative models and related experimental designs are discussed in depth. In addition, the study systematically explores the key aspects of driving simulation experiments, including scene construction, data processing, index selection and model construction, and summarizes the experimental design and data analysis methods. Finally, it looks forward to the key application directions of driving simulators in the future, such as human-machine co-driving training, function development of assisted driving system, simulation and emulation of driver behavior in the environment of vehicle-road coordination, and the construction of driving behavior model based on artificial intelligence. The research results show that driving simulators have been widely used in driver behavior research, but the authenticity and rationality of the experimental settings still need to be further improved. With the development of multi-module acquisition technology and artificial intelligence, future research needs to expand the experimental dimension and depth to capture the complexity and dynamics of driver behavior more comprehensively. Meanwhile, driving simulators still face challenges in simulating the real traffic environment, improving the accuracy of data acquisition and optimizing the model construction method, which need to be solved through interdisciplinary cooperation and technology integration.