Aiming at the problem of low distance accuracy of the harmonic distance determination method in the practical application of frequency modulation fuze, a differential frequency signal processing method is designed based on fast Fourier transform (FFT) and constant false alarm rate (CFAR) theory, with a two-dimensional fast Fourier transform (2D-FFT) method and frequency domain two-dimensional constant false alarm rate (2D-CFAR) adaptive detection method as the core. To reduce the impact of noise on the frequency modulation fusion, this approach transforms a one-dimensional differential frequency signal into a two-dimensional matrix, applies 2D-FFT transformation to produce a two-dimensional distance velocity frequency domain, and removes static clutter. The filtered two-dimensional frequency domain is used to extract the distance and velocity of the target using 2D-CFAR. The performance of the system is verified by simulation. According to the research findings, at a jamming-to-signal ratio of 15 dB, the relative error of the velocity measurement is 0.004 and the relative error of the differential frequency signal processing system is 0.033. The system has high real-time performance and can accurately output the initiation signal within 10.26 μs. The proposed method can obtain the target distance and speed accurately and in real time when the frequency modulation fuze and the ground target intersect at high speed, and improve the anti-sweep jamming ability of the frequency modulation fuze.