In the analysis of vehicular traffic flow, numerous techniques are utilized in the literature. In this thesis, distinct from the literature, to reveal the complexity of the traffic flow and its connection with the urban and traffic factors, nonextensive thermostatistics is implemented. In real systems, e.g. vehicular traffic flow, the probability distributions would become q-Gaussian and thus the use of nonextensive thermostatistics would be relevant. This approach allows a statistical interpretation to handle the given traffic flow problem. In this thesis, highway traffic flow modeling is in question in the nonextensive framework and two case studies are presented. First is related with lane changing and driver behavior, and the other is related with the superstatistics and traffic flow. In the first case study, scenario-based vehicular interactions are examined and driver behaviors are extracted by virtue of given entropy approaches. Given the configurations, Tsallis entropy approach characterizes safe driving behavior, whereas Boltzmann-Gibbs one describes unsafe driving. In the second case, vehicle speeds on the selected highway are analyzed through superstatistics theory. Two distinct q values are computed as 1.3 and 1.8 out of q-Gaussian and beta parameter distributions, respectively. The q value of 1.3 represents the highway segment with a certain flow, while the q value of 1.8 specifies the history of the traffic flow. As a result, it is revealed that the real vehicular traffic flow would involve the nonadditivity, which mainly stems from vehicular interactions as well as the urban and traffic planning decisions.