The complex electromagnetic environment is a great threat to radio fuze, taking continuous wave Doppler radio fuze for example, a method based on frequency entropy by analyzing the frequency domain characteristic fuze detection output signal is proposed. The KFCM algorithm is used for classifying and recognizing target signal and jamming signal. As the end trajectory characteristic of fuze determines the fact that the received jamming signal power increase rapidly and the signal-to-noise ratio gets worse. Thus, combined with the KFCM incremental update model, the classification model is adjusted in real time according to signal-to-noise ratio to get a better effect. The result of experiment indicate that the KFCM algorithm with incremental update has good effect on the classification and recognition of target signal at different signal-to-noise ratio, and it can effectively improve the ability of anti-jamming of radio fuze.