Abstract: Grazing angle is an important parameter when parabolic equation model is used to solve the wave propagation problem over terrain, the estimation of which has characteristic of single snapshot, coherent signals and high demand of timeliness, however, traditional estimation algorithms hardly meet the requirements simultaneously. To solve the above-mentioned problem, single snapshot spatial smoothing ESPRIT algorithm is proposed. Firstly, spatial smoothing subarrays of the single snapshot are directly used to construct the pseudo-covariance matrix to accomplish the decorrelation. Secondly, singular value decomposition (SVD) is changed into real one by unitary transformation, and estimation of signal parameters via rotational invariance techniques (ESPRIT) is instead of multiple signal classification (MUSIC) to estimate the angle. Simulation results show that the algorithm could achieve the grazing angle estimation fast and effectively, and the timeliness and accuracy could be improved compared with the existing algorithms.