Abstract: The traditional detection of wash sale in interbank market ignores the strongly planning, cooperation of wash sale and the difference between transaction subjects, and has the limitations of long running time, low efficiency and lag. In view of the problems, this paper introduced frequent itemsets and high utility itemsets to find out the trading groups who have participated in transaction together in many times, and detected the wash sale chain combining with its mode. The experiments show that the proposed method can recognize more than 97% of all wash sales and the detection time is reduced by 45%. The efficiency of detection is significantly improved. This work has a certain guiding significance for the detection of wash sale.