Abstract: A radar data storage and analysis system based on Hadoop and Spark is constructed for the problem of discovering frequent sequences from large amounts of radar data. The combination of Hadoop and Spark enables the fast storage and analysis of large amounts of radar data. In the proposed system, Hadoop uses a distributed approach to provide stable and efficient data storage, and Spark provides efficient in-memory computing. A three-node Hadoop-Spark cluster is established to execute distributed data mining algorithms. In addition, to improve the accuracy of radar data analysis, the radar data is pre-processed and the mining results are post-processed. The experimental results show that the proposed method can store and analyse large amounts of radar data efficiently and accurately.