One of the directions of aviation development is solving environmental problems, which excludes the emission of harmful substances into the atmosphere (nitric oxide, carbon monoxide) during the operation of an aircraft gas turbine engine (GTE) [1]. At low temperatures, oxygen and nitrogen are inert gases. At temperatures of 1100... 1600 K, oxides are formed, where nitrogen takes a valence of one to five. At temperatures above 1600 K, their atomic decomposition occurs. At temperatures in the range of 1100—1600 K, a reduction in NOx is possible with good mixing and a sufficient length of the combustion chamber, which determines the burning time of gases. If the combustion process is interrupted due to the poor operation of the automation, either vibro-combustion (atomic decomposition of NOx oxide) occurs at a temperature of 1600 K or flame failure occurs at 1100 K. Improving the process of converting the chemical energy of fuel and converting it into mechanical energy under conditions of uncertainty (variable caloric content of kerosene, changes in environmental parameters, wear of control equipment) is possible using neuro-fuzzy control of aviation gas turbine engine emissions into the environment. The control signal will be the fuel consumption in the diffusion manifold. In this case, fuel consumption in homogeneous reservoirs will vary evenly, provided that the total amount of fuel remains constant for the engine under consideration (the thrust should not change in the mode). A dynamic model of a neuro-fuzzy fuel consumption regulator by a diffusion collector has been developed. The method of obtaining training samples " % GT" = f (MNOx) for constructing the neural part of the regulator is presented. The desired " triangular" region of MNOx location (the integral of emission of nitrogen oxide emissions) is determined, on the basis of which control algorithms " with economy" and " without economy" of the MNOx integral are proposed.