In the pioneered Prigogine-Herman model, speed distribution change includes increase part and decrease part.For speed i, when the vehicles with speed larger than i interacts with speed i, their speed will transform to speed i and this is the increase part.When speed i interacts with the vehicles with speed smaller than i, it will transform to other speeds and this is the decrease part.This method is similar to statistical method of gases, which does not consider the anisotropy characteristic.Also, this method thinks the vehicle instantaneously decelerate,which is not consistent with the reality.In this paper, we integrate car-following behavior with Prigogine-Herman model to consider the anisotropy characteristic and driving behavior.First,we use Cell Transmission Model (CTM) to discrete the road and describe vehicle movement from cell to cell.Then, we calculate average space headway by cell density, and calculate speed difference based on speed distribution function.Further, we calibrate Wiedemann car-following model by real detector data.Last, we calculate speed distribution evolution by the calibrated Wiedemann car-following model.An numerical example will show the proposed traffic kinetic model how to remove the gas-like behavior of Prigogine-Herman model.