The interference and spoofing for satellite navigation systems pose great challenges to the safety of satellite communications. Radio fingerprint based physical layer security technique is an effective approach to solve these problems. According to the characteristics of satellite communication, this paper presents a spatial radio fingerprint extraction method based on constellation figure. The K-mean clustering algorithm is adopted to extract the radio fingerprint features related to the spatial distribution characteristics of satellites from constellation figure to identify the spoofing signals. From theoretical analysis, the spatial RF fingerprint features are extracted from the actual satellite signal and the spoofing identification is achieved based on the forged signals generated by the universal software radio peripheral (USRP) to verify the validity and practicality of the proposed method. The proposed method can identify spoofing targets without demodulating the satellite signal and can be used alone or in combination with existing spoofing target identification methods to increase the recognition rate.