Abstract: Due to the various behavior of the real driver group and different driver's ability, the acceleration decision-making and the real control decision value of the driver is not the optimal decision value. To get the difference in driver's ability and varying degrees of deviation from the optimal decision value, a multi-point and multi-objective decision-making driver model was developed based on the classic optimal preview lateral driver model, and the driving style was described with the driver's vision and decision-making intention. And the driver's ability was modeled based on the normal distribution theory. The simulink/Carsim co-simulation results show that the model including four free parameters can reflect the driving behavior of different drivers and is expected to be used to research on improving the experience of driverless car.