On ground vehicles, there is a variation of their dynamic parameters. These variations produce saturation in the actuators, even on maneuvers of moderately demand, becoming a challenge to maintain its stability and security. In order to maintain vehicle’s performance, it is important to estimate their dynamic parameters. In this paper, parameters of longitudinal and lateral dynamic of vehicles are estimated using Genetic Algorithms (GA). The GA has been applied directly in two mathematical models of vehicle, the first model is developed with scripts in MATLAB with 6 degree of freedom (DOF) and the second is the CarSim multi-degree of freedom vehicle model. The simulation in CarSim with the reference parameters, is considered as the real dynamic behavior of the vehicle. In the method of estimation, we define a vector containing the parameters to be estimated, and this evolves in each generation according to an evolution function that depends on the difference between simulated and real behavior of the vehicle. The estimated vehicle dynamic parameters are verified with different maneuvers and comparated with the parameters of reference to check the robustness of the method.