This paper investigates the synergy between Multi-Agent-based Simulation (MAS) and AI-powered queue estimation to understand the complex dynamics of skier behavior and queue formation at ski lifts. We propose a novel approach where groups of skiers are modeled as autonomous agents within the MAS framework. These agents make decisions based on factors such as preferred slopes and runs, and tolerance for waiting in lines. The simulation environment meticulously replicates the ski area layout, encompassing slopes, lifts, and other relevant features. Factors such as lift capacity and travel time are also incorporated to ensure realistic modeling. To enhance the accuracy of queue predictions, real-time data is obtained through an AI model trained on labeled images from security cameras around the ski lifts. The real-time queue length estimates derived from the AI model are then seamlessly integrated into the MAS simulation. This integration empowers the MAS with a more dynamic and data-driven understanding of queue formation throughout the day, leading to more accurate predictions of congestion patterns and wait times. By leveraging the insights gleaned from the MAS simulations, operators can make more informed decisions regarding lift operations and strategic layout modifications. Optimized lift usage based on these predictions can potentially lead to reduced waiting times for skiers, significantly enhancing their overall experience. In conclusion, this paper presents a powerful framework that combines MAS and AI-powered queue estimation to offer ski resorts a valuable tool for optimizing skier experience and operational efficiency.