<p>Regular detection of highway tunnels is an important task in highway maintenance, and it can detect tunnel disease timely. However, the current evaluation standards for highway tunnel detection data are single, which cannot reflect the correlation between tunnel material units and the degree of correlation between diseases over long-term operation. These data require deep mining to explore their disease patterns. Therefore, a dynamic clustering matrix analysis method is proposed to address the issues of tunnel disease correlation and development patterns in long-term detection based on a highway tunnel as background. The method uses a quadratic approach to solve the transfer closure fuzzy matrix and analyzes the truncation matrix to obtain the correlation eigenvalues of the dynamic clustering matrix. This method is suitable for analyzing the correlation between different disease units and types in tunnels. The results indicate that there is a certain time series correlation between tunnel disease types, and the distribution of diseases among different tunnel disease units also shows a certain spatial distribution correlation. In the early stage of the disease, circumferential cracks have a greater impact on water leakage and damage, while longitudinal cracks have a greater impact on water leakage and damage in the later stage of the disease. In terms of the degree of disease development, circumferential cracks and water leakage have a greater impact, followed by longitudinal cracks, and the damage has the smallest impact. This study can facilitate in-depth analysis of long-term tunnel disease detection data, grasp the dynamic evolution law of tunnel diseases, and is of great significance for timely maintenance decision-making.</p>