<p>Seismic semblance attributes are important for highlighting geological edge structures, such as faults, in seismic coherence technology. It relies on the similarity of waveforms of adjacent seismic traces to achieve discontinuity detection. However, traditional methods have limited adaptability and often do not meet the requirements for detecting complex underground structural edges. We systematically extend the conventional semblance attribute method and innovatively propose the block semblance attribute algorithm, which reconfigures the traditional semblance calculation into an optimization problem-solving process, providing the semblance attribute with a new mathematical interpretation. By adjusting the operation matrix in the objective function, various targeted attributes (e.g., dip average, Gabor filtering, gradient, and symmetric block semblances) can be derived to meet diverse geological edge detection needs. This adaptability allows for targeted logic optimization: symmetric block semblance excludes layer-edge interference by removing event axis constraints to accurately identify faults, while Gabor filtering block semblance retains channel edge information by applying local dip angle constraints to highlight reflection displacement. The actual seismic data processing results reveal that block semblance attributes can effectively obtain abundant boundary information, have good detection effects on complex fault edges and layer-edge river channels, and can improve the seismic edge detection attribute system, providing more accurate support for the interpretation of complex geological bodies.</p>