For peanut production is essential that the harvest follow the same alignment of the sowing, which avoids probable positioning errors during the operation, which can lead to high levels of losses. This problem can be avoided when using automatic orientation, however, little is known of the effectiveness of satellite signals (RTX) in directing the machines in the field. The goal this paper was to quantify and analyze the probability of positioning errors in peanut sowing using automatic orientation, RTX signal, in two paths, curved and rectilinear. Parallelism error and execution error were monitored during 12 passes of the mechanized set in each path. The error was calculated by the orthogonal difference between the projected and the executed path, and then the probability distribution and the occurrence frequency of the errors were performed. Most errors between the passes were 8.75 and 5.22 cm in the curved and rectilinear path, respectively. The probability and frequency of occurrence of errors greater than 4 cm are inversely proportional when using RTX signal in peanut sowing. Most errors found in both trajectories do not interfere with the overlapping rate since they are less than 0.9 m, recommended spacing between the rows of the crop