Abstract: Traffic accidents have become a pressing societal issue,endangering lives and property.The Internet of Vehicles (IoV) technology has the potential to significantly improve road safety.This paper proposes a novel two-player stochastic game model,grounded in Markov decision processes,to aid IoV system operators in understanding the dynamic interactions between attackers and defenders,as well as the cascading effects between different system states.The attacked IoV system is divided into three phases:penetration,disruption,and proactive recovery.The deployment cost of IoV system is divided into intertwined cyber and physical layers.By refining the cost settings for both attackers and defenders,Minimax-Q is employed to predict attacker actions,even with incomplete knowledge of model parameters.The impact of defense effectiveness and cybersecurity deployment cost on attacker strategies is explored,assessing the system reliability under various attack strategies.Key metrics include Vehicle System Mean Time to Failure (MTTF),Vehicle System Steady-State Availability (SSA),and Road Information Steady-State Confidentiality (SSC).A novel composite index,the Road Safety and Traffic Integrity Indicator (RTSL),is formulated to assess how IoV system cybersecurity reliability influences road traffic safety.Our findings indicate:(1)In the penetration phase,fluctuations in the deployment cost of the physical layer’s security have a more significant influence on attack strategies compared to the cyber layer,while the trend reverses in the disruption phase.(2)Under a specific compromised scenario A,the IoV system’s robust adaptability enables rapid detection and mitigation of increased attack likelihood,leading to improvements of 3.9% in SSA,3.9% in SSC,and 1.34% in RTSL.(3) The penetration phase (N,V,P) is identified as a crucial period affecting the IoV system’s impact on road traffic safety,posing a significant threat to road traffic safety.These insights assist IoV system operators in making informed cybersecurity resource allocations,improving cybersecurity in critical states,and ultimately the overall road traffic safety level.