Absrtact: In order to ensure the safety, stability and driving comfort of vehicle lane keeping in low-speed scenarios, this paper proposes an evaluation method of lane keeping behavior capability based on automatic driving technology. Taking low-speed scene as the research object, based on automatic driving technology, IDM intelligent driving model is applied to describe ACC car following model, and MATLAB/Simulink and Carsim are used to build a low-speed scene lane keeping simulation platform The three aspects of comfort are used to evaluate the vehicle lane keeping ability in low-speed scenarios. The structure shows that when the distance between the target vehicle behind and the vehicle in front reaches a stable safety distance, the average speed of the vehicle is 21.6 (km · h -1 ), and the floating size is not more than 1. The difference between floating up and floating down is 0.0002, which is conducive to the stability of the lane during driving. In addition, based on the IDM intelligent driving model, this paper obtains data through vehicle to vehicle communication in the urban low-speed scenario, that is, when the vehicle is driving in the lane, its acceleration changes slightly. When it approaches zero, it can reduce unnecessary deceleration, so as to improve the driver's riding comfort. Based on ACC simulation, the difference between the actual following distance between the target vehicle and the vehicle in front during lane keeping and the target following distance fluctuates between -0.06 and 0.05m, and most of the differences are greater than the average difference, which can well meet the actual requirements and make car following safer. It can be seen that it is effective to study the lane keeping ability of vehicles in low-speed scenarios based on automatic driving technology in this paper.