Adaptive beam-forming technology based on linear constrained minimum variance linear constrained minimum variance（LCMV） criterion algorithm can ensure the maximum signal gain in the desired direction and the null formation in undesired direction in ideal condition.In practical array systems

traditional cirterion algorithm based on the LCMV are known to degrade if there exits pointing error

sensor positioning error

and sensor phase error.An improved SD-LCMV algorithm based on Newton method and variable loading constraint is proposed to solve the problem of optimal weight vectors resolving. Simulations of two algorithms are compared with linear constrained LMS algorithm and recursive robust LCMV algorithm. The results verify robustness of the improved algorithm to error.