Abstarct: Abstract The protection of human safety and electrical equipment against unwanted voltages and electrocution in power systems is entrusted to a set of conductors that are placed in the ground and are called the ground network. The high importance of the ground network in equipment protection has made the modeling and calculation of the impedance of the ground network and not exceeding the standard value become one of the challenges in this field. Currently, various methods are used for this work, including circuit methods and transmission line methods, which use simplifying approximations and assumptions, and the electromagnetic method, which performs calculations with accurate solutions. In the first two methods, due to the use of approximation, the accuracy of the results decreases, and in the third method, we face heavy and time-consuming calculations. In this thesis, using artificial intelligence, machine learning, and more precisely, the Gaussian process model algorithm, we presented a new method that can calculate the earth's impedance in a much shorter period of time and by bypassing heavy calculations. Also, in this method, the direct effect of various variables such as wire diameter, burial depth, soil permeability coefficient, etc. is considered. Our proposed algorithm provides a function for predicting ground impedance values in terms of frequency and other variables with a very good speed compared to previous methods using a databaxse obtained from exact computing methods (MOM).