Abstract: Over-the-horizon radar（OTHR） has rather wide scope of monitoring，which is often required to detect ship target under the conditions of short coherence integration time（CIT）in practice，however，short CIT causes low Doppler resolution，leading to the difficulty of detecting ship targets from the strong sea clutter.To solve the above problem，sea clutter suppression algorithm of subspace estimation based on higher order singular value decomposition（HOSVD） is proposed.Firstly，Hankel tensor is established by echo data of the distance unit containing the ship target.Then the sea clutter subspace and target subspace of that tensor are solved by HOSVD.Finally，in order to suppress sea clutter，Hankel tensor is mapped to the target subspace by orthogonal projection method.It shows in the simulation results that the signal to clutter plus noise ratio（SCNR）and the peak sidelobe level ratio（PSLR） are enhanced，compared with the existiing sea clutter suppression method of subspace.