To solve a number of problems of control and diagnostics of gas turbine engines, dynamic models are used, the coefficients of which in a number of practically important cases must be determined from real information obtained experimentally or during operation. The paper proposes a method for analyzing errors in estimating the dynamic parameters of engines, the main concepts of which consist in using a linear dynamic model, considering influencing errors as independent normally distributed random variables, using dimensionless parameters and the corresponding transformation of equations that describe the relationship between estimation errors and influencing factors, which makes these equations universal, suitable for analyzing any parameters of the engine operating process and any control actions. It is established that, depending on the formation of errors in the initial data, they can be divided into two categories: parameters distributed over a set of experiments, and in each experiment – over a set of local measurements, as well as parameters distributed only over a set of experiments, the values of which are fixed over a set of local measurements. Using analytical and numerical methods, universal dependencies have been obtained that make it possible to determine the errors in estimating the time constant of a single-shaft gas generator with a sudden change in fuel consumption and with a linear change with stabilization depending on the time and frequency of observation, the expected value of the time constant, as well as the root-mean-square errors of the initial data: measurements of the rotor speed and fuel consumption, as well as the gain factor for fuel supply.