To clarify the safety of mixed traffic flow on highways under foggy conditions, this study conducted a systematic simulation analysis of mixed traffic flow in risk scenarios. By examining the information perception patterns of human-driven vehicles (HDVs), automated vehicles (AVs), and intelligent connected vehicles (ICVs), this study developed corresponding car-following models. Simulation scenarios with varying risk levels were created, and the simulation data were analyzed using indicators such as the cumulative reciprocal of time to collision (TTC), average speed, and speed standard deviation. The results indicate that, compared to HDVs, non-HDV vehicles (AVs and ICVs) significantly enhance the safety and stability of mixed traffic flow. In high-risk scenarios (deceleration of 6 m/s2), the cumulative reciprocal of TTC for ICVs is 12.8% lower than that for AVs. Furthermore, ICVs better maintain traffic flow stability than AVs, with the speed standard deviation in high-risk scenarios being reduced by an extra 13.1% for ICVs. Notably, ICVs have even shorter perception reaction times and display smaller speed adjustment oscillations. The findings of this study offer valuable insights for managing mixed traffic flow on highways during foggy conditions in the future.