The invention relates to a high-dimensional S transforming method based on structure tensor guiding. The high-dimensional S transforming method comprises the following steps that a structure tensor matrix is constructed, and seismic data are subjected to anisotropy parameter extraction through the feature value and the feature vector of the structure tensor matrix; a high-dimensional S transforming operator based on structure tensor guiding is constructed; and the seismic data are subjected to high-dimensional S transforming through the high-dimensional S transforming operator based on structure tensor guiding, and sparse representing data according with the feature of the seismic data are obtained. The seismic data can be subjected to improved high-dimensional S transforming according tofeature information, analyzed by the structure tensor, of the seismic data, and thus the seismic data are more prominent in feature, refined and sparse; and the information utilization rate of the seismic data during processing and interpretation is increased, extraction of the property of the seismic data is facilitated, the important theoretical and application value is achieved, and the certainpromoting effect on development of oil gas and mineral resource exploration is achieved.