Abstract – Traditional identification algorithms that are based on model parameter adjustment often have methodical errors. These errors prevent a practical application of identification results in automated computer control and diagnosis of technical objects. Methodical errors have a hidden character and, consequently, software stops working. The obtained numerical data have abstract character and no reasonable physical interpretation. There are not any mathematical methods that could be applied to decode them. Such conclusion follows from the proofs obtained in an analytical form in this paper. Until recently, they could not be obtained using traditional computing methods. It required finding, in an analytical form, the solution of systems of difference equations on the basis of which identification models are formed. With the help of the method described in this paper, this problem has been solved. It is proved that introducing into the model any additional operators, as it is done in stochastic models [17], [18], is undesirable. Also it is proved that it is necessary to carry out full cycle of identification, instead of stopping it at the stage of estimating the discrepancy of the system of difference equations. The process should end with checking the reliability of estimations of parameters of analog object’s transfer function.