Abstarct: Faults play a significant role in the exploration, extraction, and exploitation of hydrocarbon resources. One of the most important and widely used tools for identifying these geological events is seismic data. Although faults cause discontinuities in seismic data, enabling their identification, due to various factors such as the destructive impact of noise and the time-consuming nature of manual interpretation, seismic attributes are now employed for fault detection and identification in seismic data. Various attributes have been introduced for detecting seismic discontinuities, among which coherence is one of the most common and widely used attributes. In this study, a novel coherence attribute baxsed on eigenvectors and eigenvalues of the covariance matrix is introduced and examined. Compared to the conventional eigenstructure-baxsed coherence attribute, this method utilizes both eigenvectors and eigenvalues for coherence analysis. The inclusion of eigenvectors enhances the method's sensitivity to changes in shape and polarization, while purely eigenvalue-baxsed analysis is more sensitive to energy variations. The results of evaluating the proposed attribute on synthetic and real data demonstrate its successful performance in identifying discontinuities caused by reflector displacements. Moreover, it provides higher resolution compared to conventional semblance and eigenstructure-baxsed methods. Additionally, the proposed attribute exhibits superior performance in identifying discontinuities related to polarization changes and wavelet deformation. Furthermore, the analysis of results shows that the improved eigenstructure-baxsed coherence attribute is more robust against noise compared to semblance -baxsed methods. These features make the proposed method a more efficient tool for fault and discontinuity detection in seismic data.