Connected and autonomous vehicles (CAVs) are under increasing cyber-attacks from multiple threat-agent types, such as nation-states, cybercriminals, terrorists, hacktivists, etc. Each threat-agent type varies in its capability, available resources, valuation of target, and strategies for successful attack. On the other hand, the defender has limited resources that need to be allocated effectively and efficiently to mitigate and prevent cyber-attacks in CAVs when faced with adaptive attackers and attack strategies. This paper proposes a novel class of game theoretic model to study the multi-target attacker-defender game under limited budget constraints. The proposed game model investigates the change in the defender's and attacker's equilibrium resource allocation considering the budget available, valuation of the target, cost-effectiveness, and implementation cost of attacker and defender in a sequential move game. The objective of each attacker is to maximize the total expected loss, and the defender aims to minimize the total expected loss by distributing a limited amount of resources to multiple targets. The analysis shows that a higher amount of resources should be allocated to the most valuable target. The sensitivity analysis shows that the defender’s expected loss and the attacker’s expected profit increase with an increase in the attacker’s budget. With an increase in the defender’s budget, the defender’s expected loss and attacker’s expected profit decrease.