With the popularity of electric vehicles, the charging demand from users continues to rise. Efficient charging scheduling is important for improving overall efficiency and alleviating the burden on the power grid. However, due to limited spectrum resources, uploading data collected by large-scale charging piles to base stations (BSs) may intensify channel competition and increase transmission delay. These issues affect scheduling efficiency and prolong user waiting time. To address this, we propose a D2D clustering resource allocation method for charging pile information collection. In order to maximize system throughput, we establish a spectrum resource allocation model based on D2D communication for charging pile clusters. We employ a genetic algorithm (GA) to optimize the channel allocation for cluster heads and intra-cluster members. Simulation results show that this method can effectively alleviate channel competition, improve system throughput, and provide better communication support for intelligent charging scheduling.