The relevance of the work is caused by the need to develop the algorithmic methods of improving reliability for automatic control systems of new generation of gas-turbine engines (ACS GTE). The ACS GTE are used in different branches of industry, including gas-turbine power plants utilizing oil-associated gas on oil fields of the Russian Federation. The main aim of the study is to evaluate the effectiveness of developing algorithmic redundancy using the built-in mathematical models to improve the ACS GTE reliability; to develop the robust adaptive linear gas-turbine engine model, operating in real time, satisfying the high accuracy requirements under deterministic, stochastic and random changes in the engine condition. The methods used in the study. The dynamic linear model is built by the state space method; herewith the individual engine throttle characteristic is used as the basis of static model. This throttle characteristic was obtained in the acceptance tests or "race" during the operation after the maintenance. The account of the deterministic changes in the engine condition is based on the method of adjusting the static characteristics to the worsening of engine components efficiency due to expenditure of resources, which is defined by a matrix of influence coefficients. The account of the stochastic variations is based on the method of analysis of the diagnostic matrix of situations, including the use of numerical methods Mont Carlo using the random numbers sequences obtained by the I. M. Sobol algorithm (LP-sequences). The account of the random changes is based on the methods of one-dimensional and multi-dimensional Kalman-filtering. The methods of object-oriented programming in C ++ and experimental modeling in MatLab were used for developing the algorithms. The results. The authors have developed the algorithms of the adaptive built-in mathematical model of an gas-turbine engine which allows determining the failure of information channels, and replacing the failed channel information in case of deterministic, stochastic and random changes of the engine parameters. The carried out non-motorized and motorized bench developed algorithms' testing showed their operability and high efficiency to improve the ACS GTE reliability.