Abstract: Vehicle networking is a hot spot in the world transportation industry. Vehicle networking can become more intelligent by processing and mining new information in these data in real time. With the development of Vehicle networking, the massive data collected from device monitoring requires higher data storage and query performance. This paper uses Apache Spark as a unified cluster computing platform based on HDFS high-availability clusters, proposes and designs a Spark-based vehicle trajectory data analysis method, and uses area division to mine large amounts of data to solve road congestion problems and to assist urban roads. Planning has certain research significance.