Abstract: In order to conduct precision-distribution more precisely, it is necessary to make use of contemporary theories of error synthesis, among which the algorithm of multiple convolution is especially efficient. Yet it necessitates fairly large amounts of calculation, takes much time and thus becomes rather complicated when applied to problems in precision - distribution. In order to solve this problem satisfactorily, it is proposed to settle the factors beforehand on the basis of various principles for precision-distribution. With this complicated calculations in the process of repeated optimization can be prevented, and problems of precision-distribution can be conveniently and more promptly solved, attaining eventually accurate results.