To address the challenge of quantitative thickness measurement for thin ice layers during the initial aircraft icing phase caused by overlapping ultrasonic echo signals， this study presents a method to inversely determine thin ice thickness using the resonance frequency of the ultrasonic reflection coefficient amplitude spectrum （URCAS） at the skin-ice interface. Numerical simulations are conducted to obtain raw ultrasonic echo signals from aluminum skin surfaces during early icing， followed by URCAS calculations. The upper and lower measurement limits of ice thickness are determined based on the -6 dB bandwidth of the incident wave amplitude spectrum. Advanced signal processing techniques are developed to extract critical time-frequency features from theoretical echo signals. The resonance frequency points in URCAS are specifically utilized for ice thickness inversion. The URCAS-based method enables effective extraction of characteristic parameters from overlapping ultrasonic signals during initial icing. The lower measurement limit of ice thickness is approximately half of that achieved by conventional time-domain analysis. Theoretical relative measurement errors remain below 4%.