To address the multi-objective travel route planning problem, which integrates tourists' personalized preferences, real-time congestion levels of scenic spots, and commuting time, an interactive travel plan intelligent generation planner based on a large language model (LLM) is proposed. First, a LLM information processing module identifys, reasons, and transforms user requirements into structured data. Then, a scenic spot tourist flow prediction module based on the random forest algorithm is constructed, which integrates multi-dimensional factors such as historical tourist flow, weather, and holidays to achieve accurate tourist flow prediction. The prediction results are mapped to congestion levels through the LLM information processing module. Finally, a multi-objective travel route planner centered on the parallel non-dominated sorting genetic algorithm-II (PNSGA-II) is built to achieve intelligent route planning and search for global optimal solutions. The simulation experimental results indicate that the PNSGA-II outperforms other multi-objective optimization algorithms in terms of solution quality and computational efficiency. Additionally, this planner demonstrates significant advantages in both plan effectiveness and generation efficiency compared to others.