The study covers the development of a mathematical model of a micro gas turbine (MGTE) operating under transient conditions using a recurrent neural network. The compressor inlet temperature and pressure depending on the aircraft height and speed are taken into account explicitly. A full-size mathematical dynamic MGTE model based on engine per-unit description was used to verify the developed model. The obtained model was compared with the existing one employing normalized parameters of aircraft flight level and airspeed. The simulation suggests that the proposed model yields significantly smaller errors than the existing one, whereas the computation time of both models differs insignificantly.