Frequency-scanning interferometry (FSI) is a high-precision distance measurement method. When applied to high-speed moving targets in narrow spaces, the initial short distance combined with significant Doppler effects due to motion can cause the FSI signal to exhibit negative frequency components. Because frequency identification methods only recognize positive frequencies, phase extraction is limited to phase increments, which in turn leads to errors in the initial clearance value derived from the slope demodulation of the phase increment. An axial clearance demodulation algorithm is proposed for FSI signals, based on negative frequency identification and phase correction. By extracting time-frequency signals, using a bidirectional threshold to determine the zero-crossing point of the frequency, and constructing a sign judgment square wave function, the method corrects the frequency interval and the relative phase variation, thereby overcoming the inherent limitations of FSI that require a large initial clearance and low rotational speed. Experimental results demonstrate that at a rotational speed of 12,000 rpm, the corrected average clearance relative to the static average clearance reduces the error from 36.5% to 0.4%.