In our time, problems of accurate positioning of objects, equipment and cargo arise more and more often. These are the tasks of logistics and precision farming and, a separate group of tasks, positioning during earthworks in construction: pits, arrangement of pile fields, etc. If a positioning accuracy of 10 m is quite sufficient for logistics tasks, monitoring the movement of transport and cargo, then for construction tasks, an accuracy of 10 cm is required, and building monitoring tasks require a millimeter accuracy. This work is devoted to increasing the accuracy of satellite geolocation by reducing the random error of the receiver. A budget receiver with a price of about $5 was chosen for the study, from which data on the current geolocation was recorded. The data from the receiver contains a random error, which was eliminated by averaging the result. Studies have shown that with small averaging intervals, the error remains, although it decreases; as the averaging interval increases, the error decreases, however, up to a certain width of the interval; a further increase in the averaging interval ceases to affect the magnitude of the error. It is optimal to average the results in the interval from 1000 to 2000.