The number of spoofing attacks is more noticeable with the increasing use of navigation and positioning systems and with the apperance of inexpensive software-defined radios. In order to detect and avoid such attacks at the right time, it is necessary to work on improving systems for detecting these spoofing attacks and suppress such attacks as much as possible. There are many different spoofing detection methods which can be hardware or software even machine learning methods. The GNSS system with features and the protocol which it uses for communication are described in this paper. The spoofing attack is described in more detail, as well as various methods for spoofing detection. The practical part of the work includes development of a laboratory model for performing spoofing attack with the application GNSS-SDR which is used to collect location data. Using Classification Learner application, the GNSS signal classification was carried out with the purpose of recognizing spoofed signal. It is shown how different machine learning models classify signals into classes real and spoofed with high accuracy. The high accuracy of the trained models confirms the conclusion that these methods are suitable and reliable approach for spoofed signal detection in the GNSS system.