in this paper the authors propose a general approach to developing neural network algorithms aimed at monitoring of signal parameters and have developed a neural network algorithm designed to monitor the absolute value of a linear chirp (LFM) signal phase deviation. The authors show some possible ways of how to reduce an error probability of the neural algorithms when monitoring signal parameters in the technical check and control systems.