This paper discusses a method to reduce calculation load by means of layered searching construction of association hypothesis events. The layer here can take 0 value or an integer L. Layer 0 means that no target can find its association data from current returns. Layer L means that L targets can find their association data from current returns. Our layered construction method in the paper is recursive and of independence among layers, so it can also be implemented in parallel structure. Comparative analysis of the method with relative methods in other references and the corresponding computer simulation tests and the results are also given in the paper.