Aiming at the problems of heavy workload and unstable switching process caused by the design algorithm of separate working conditions in the existing longitudinal control strategy, this paper proposes a unified preview decision-making model, which imitates human driving behavior, and comprehensively decides the optimal longitudinal acceleration through safety, legality, and evaluation indicators; in addition, for the safe distance model based on kinematics analysis that has poor traffic efficiency and does not consider the driver’s driving habits, this paper proposes a safe distance model based on the driver’s preview characteristics and a safety judgment method based on predicted lanes; the human-vehicle-road closed-loop simulation test shows that the preview decision-making proposed in this paper can meet the expected acceleration decision of various traffic conditions. The safety distance model and safety judgment method based on the preview characteristics are scientific, reasonable, and effective.