Introduction. Nowadays, communication systems are mostly digital. One of the tasks of automatic gain control in digital receivers is to maintain analog signals at the appropriately fixed level, which prevents saturation of the analogto-digital converter. Most numerical algorithms are based on floating point arithmetic, and digital automatic gain control is usually implemented using fixed-point arithmetic devices such as programmable logic chips and signal processors. As consequence of fixed-point arithmetic and hardware constraints usage, the out-put significant bits should be truncated correctly. Although many studies mention digital automatic gain control, its characteristics are not considered in detail in terms of the finite capacity of calculators.Objective. The purpose of the study is to analyze dynamic characteristics of digital automatic gain control implemented on a computer for operations on numbers with fixed-point.Materials and methods. Within the frames of the study in Matlab software was developed a mathematical model of digital automatic gain control. The model was implemented on a programmable logic chip.Results. The paper shows the difference in characteristics and features of the digital automatic gain control during operations on fixed-point numbers. The study provides the assessment of the effect of fixed-point signals on the stability of the digital automatic gain control and includes the analysis of causes of spurious oscillations of the control signal.Conclusion. The study proposes the algorithm for compensation of the control signal oscillations by means of correction of the reference level of the digital automatic gain control. Further is required to verify the proposed algorithm on real signals. The results of the study are relevant in development of digital receivers for communication systems of various purposes.