<p>The seismic discontinuity attribute body can reflect the spatial characteristics of faults to a certain extent. The direct use of the discontinuity attribute body to identify faults is often susceptible to noise interference and has poor detection performance for complex faults. The conventional fault enhancement processing algorithm is constrained by fault direction information, processes discontinuous attribute bodies, and suppresses noise interference in non-fault directions. Its calculation accuracy depends on the calculation accuracy of the fault direction, and there are significant errors at the intersection of faults. Therefore, based on the distribution pattern of local linear features of faults in two-dimensional cross-sections and local planar features in three-dimensional space, a fault enhancement method based on multidirectional component analysis is proposed. In the method, the local fault signals of discontinuous attribute data are treated as a mixture of plane wave signals from multiple different directions, and the mixed signal is decomposed into multiple unidirectional plane wave signals using multidirectional component analysis. The calculation process of the multidirectional component analysis algorithm itself can achieve denoising, and the obtained multiple single-directional plane wave signals are smoothed to enhance the single-directional plane waves. The results of the synthesis of seismic data through fault models and the processing of actual seismic data reveal that the fault enhancement method based on multidirectional component analysis can effectively suppress noise in complex fault structures, enhance the continuity of faults, and improve the clarity of faults, especially at the intersection of local faults where the enhancement effect is significant.</p>