Abstract: The signal of the skywave over-the-horizon radar is reflected by the ionosphere. The perturbation of the ionosphere modulates the phase of the backscattered radar signal, and spread its spectrum. The spread clutter masks the target and makes it difficult to be detected. Before target detection is performed, in order to improve the probability of detection, it is necessary to correct the contaminated radar return. Instantaneous frequency estimation plays a key role in decontamination. The return signal is modeled using a cubic phase (CP) component, and a novel frequency estimation algorithm is proposed based on this model. The simulation result shows that comparing with the Hankel rank reduction(HRR) algorithm, the new algorithm can track larger frequency perturbations thus has better decontamination performance.