A new approach is proposed to obtain a generalized model of distribu -ted digital fiber-optic measuring systems of interferometric type using multichannel reception of signals of a fiber-optic inter-mode interferometer to improve the accuracy of measurements. On the basis of this approach, generalized equations for the con-version of fiber-to-digital converters of the geometric coordinates of the points of the measured object are obtained. The equations combine all the private mathema ti cal models of energy information processes. The approach is based on the representa-tion of the "coordinate of point (move) — code" in the form of an equation of perfect digital-to-analog source code conversion, the processes of which change bit codes are given in the form of logical functions from the input move and points of real multidimensional spatial parameters. The fiber optic line is used in bidirectional optical sig-nal mode in conjunction with the code element element. In this function, the supply of radiation from the measuring units to the points of reading information, the control ele -ment, transmitters of modulated radiation are combined in a single fiber. The spatial separation of optical streams is carried out in a block of bidirectional optical communication devices, which is a set of fiber-optic Y-splitters. For multichannel reception, the principle of making a decision on registration of influence on the interferometer is in-troduced: if the module of the output signal exceeds the set level, the signal is fixed. Changes in the measuring signal from external conditions are determined by changes in the parameters of the fiber, the processes of interaction of modes and double re-fraction. Changes in the measurement signal are presented as random variables. Using the central limit theorem for a large number of double sums, the values of the signals at a particular point in time are described by independent random variables, with a normal distribution law and a variance. The beneficial effect is considered regu-lar, and at the time of measurement it is represented by a centered Gaussian random variable with variance. The useful signal component is a Gaussian random variable with standard deviation.