Abstract: QR decomposition LS algorithm (QRD) is a widely used algorithm for adaptive arrays recently,which has good numerical stability and can be easily mapped to Systolic structure for parallel processing.However,for full adaptive array where optimal weights are needed,this method typically requries three Systolic modulars to produce optimal weight.This paper presents a new method to solve the nulling problem of a full adaptive array using inverse QR decomposition LS algorithm (IQRD) and its Systolic implementation.It can easily obtain optimal weight vector using only one Systolic array.Since it still directly deals with the input data matrix and is based on unitary transformation,the IQRD algorithm has the same numerical stability as that of the QRD algorithm.