As one of the critical technology of cognitive radio, spectrum sensing detects the surrounding electromagnetic environment in real-time to make sure whether ‘spectrum hole’ exists and primary users’ spectrum. Low earth orbit (LEO) satellite earth station is a kind of typical single user perception nodes, and it has the characteristics of fixed geographical location and spectrum environment. However, with its antenna elevation changing, the received signal SNR is not fixed. The spectrum environment around the earth station needs to be sensed for improving the spectrum utilization efficiency of satellite earth station. According to the characteristics of multi-resolution detection algorithm with fast detection speed and matched filter detection algorithm with high detection probability, a two-step detection algorithm is designed based on the corporation of multi-resolution detection and matched filter detection. Through theoretical analysis and experimental simulation and comparing with energy detection, this method shortens the detection time and improves the accuracy of detection.