Evacuation systems play a critical role in mitigating casualties and property damage during emergency situations. Comprehending evacuee behavior in overcrowded scenarios is vital for developing effective evacuation strategies. Nonetheless, evacuating large crowds from buildings with multiple exits presents a significant challenge, particularly when exits are asymmetrical and crowds are dense. This paper proposes an innovative two-dimensional cellular automaton model for multi-exit evacuation, designed to streamline evacuee decision-making in an asymmetrical exit layout within dense crowds. The proposed model addresses limitations in traditional evacuation models by incorporating dynamic counting areas and enhanced parameters, including congestion, distance, and empty parameters