Existing studies on the Autonomous Emergency Braking (AEB) system have mostly centered on forward collision avoidance. Since they do not incorporate motion parameters of the rear vehicle, such systems tend to apply excessive braking, which in turn increases the risk of being rear-ended by the rear vehicle. To address this issue, this paper proposes an AEB control strategy that takes into account the states of both the front and rear vehicles. The strategy involves constructing a critical deceleration model for rear-end collision prevention and determining the host vehicle’s deceleration model based on forward collision risk assessment. Meanwhile, the variable trigger threshold of the AEB system is defined using the aforementioned models. Specifically, when a forward collision risk is detected, the system prioritizes braking with the deceleration required for rear-end collision avoidance. If this deceleration fails to prevent a forward collision, it switches to the minimum deceleration needed to avoid the forward collision, thereby minimizing the risk of being rear-ended by the rear vehicle. Simulation comparisons and verification under typical operating conditions show that, compared with the traditional TTC (Time-to-Collision) model, this algorithm can fully consider the motion state of the rear vehicle, balance collision risks from both the front and rear, and effectively avoid the dual safety hazards of forward collision and being rear-ended by the rear vehicle.