【目的】 在“双碳”背景下，我国电动汽车保有量迅速增加。然而，现有电动汽车充电设施的建设速度未能满足其快速增长的充电需求，高速公路服务区充电困难和排队拥堵问题尤为突出。 【方法】 首先，基于地理信息系统数据和实际交通流量，构建了出行概率矩阵，并引入充电焦虑系数，以增强模型对真实场景的适应能力。其次，提出“车流量对车速的影响系数”，进一步提高了车辆到达服务区时电池荷电状态的预测精度，并据此预测高速公路服务区的充电负荷。最后，基于预测结果建立了多服务区充电桩优化配置模型。该模型引入了基于车流量的动态权重调整机制，通过动态分配建设成本、运维成本和用户等待时间成本的权重，实现了在不同交通负荷条件下优化目标权重系数的灵活调整。 【结果】 仿真结果表明，充电桩数量的提升虽显著降低用户等待时间，但伴随投资成本上升；节假日期间充电负荷峰值较常态提升26.3%，呈“双峰”特征，且需求集中于行程后半段服务区；所提出的动态权重调整机制可随车流波动，实现成本与用户体验的自适应平衡。 【结论】通过引入动态权重调整机制，对多服务区充电桩配置进行优化，显著提升了高速公路充电网络在供需匹配中的灵活性与响应性；在车流波动与节假日需求峰值的双重作用下，实现了充电资源与用户需求的精准耦合，不仅提升了桩位利用率、缩短了排队时间，同时有效协调了建设与运维成本。

[Objective] To achieve carbon-peaking and -neutrality goals， the number of electric vehicles （EVs） in China has grown rapidly. However， the growth of EV-charging infrastructure has lagged behind the rapid increase in demand. The problem of inadequate charging capacity is particularly severe in highway service areas， where challenges such as difficulty in accessing chargers and long waiting queues have become increasingly pronounced. [Methods] First， based on a geographic information system data and actual traffic flow， a travel probability matrix is constructed and a “charging-anxiety coefficient” is introduced to enhance the model’s adaptability to real-world scenarios. Second， a “traffic-flow-to-speed impact coefficient” is proposed to improve the accuracy of predicting the state of charge of EVs upon arrival at highway service areas； this coefficient serves as the foundation for forecasting charging loads. Finally， a multi-service-area charging-pile optimization model is developed based on the forecast results. This model incorporates a dynamic weight adjustment mechanism based on traffic flow， allowing for the flexible reallocation of construction， operation and maintenance， and user-waiting-time costs under varied traffic load conditions， thereby enabling adaptive optimization of objective function weights. [Results] Simulation results show that although increasing the number of charging piles significantly reduces user waiting times， it also leads to higher investment costs. Peak charging loads during holidays increase by 26.3% compared with those during normal periods， exhibiting a bimodal pattern， with demand concentrated in service areas located in the last half of travel routes. The proposed dynamic weight adjustment mechanism adapts to traffic flow fluctuations， enabling an adaptive balance between costs and user experience. [Conclusions] Introducing a dynamic weight adjustment mechanism and optimizing charging-pile allocation across multiple service areas significantly enhance the flexibility and responsiveness of highway charging networks in matching supply with demand. Under the combined effects of traffic flow fluctuations and holiday demand peaks， the proposed approach achieves precise coupling between charging resources and user needs. This not only improves pile utilization rates and reduces queuing times but also effectively balances construction and operational costs.