For the problems such as the lack of tire braking force, the wheel locking-up and the increase of articulated forces between the vehicles caused by the change of road adhesion condition, an automatic emergency braking strategy of the virtual rail train, which is a new multi-axle articulated urban rail transportation vehicle, is proposed considering road adhesion coefficient. The proposed strategy consists of two-layer controllers. Firstly, the upper controller is designed by extended Kalman filter and model predictive control, and the desired braking acceleration of the train is determined according to the real-time estimated road adhesion coefficient and the braking performances of comfort and safety. Secondly, the lower controller is designed by the principles of load proportion distribution and PID feedback compensation control, and the braking torques of train tires are determined based on the desired braking acceleration of the train. Finally, the cosimulation platform of Matlab/Simulink and Trucksim is built to research the vehicle articulated forces of virtual rail train under emergency braking conditions. The results show that the variations of train load, initial braking speed of the train and road adhesion coefficient affect the maximum articulated force between the vehicles. The proposed strategy processes good robustness. It can prevent the wheels locking-up, reduce the maximum articulated force between the vehicles and improve the braking performances of comfort and safety under various conditions.