<p>Panel data can well describe and depict the systemic and dynamic of the research objects. However, it is difficult for traditional panel data analysis methods to accurately reflect the structural information of the research object and scientifically deal with the multi-attribute decision problem of presenting the similarity of systemic structure among the research objects. In view of this, with respect to the clustering problems for panel data, considering system structural characteristics of the research objects such as scale volume, component weight, development trend and volatility, by using theories and methods such as GM(1,1) and grey incidence clustering method, we construct a grey structure incidence clustering method for panel data, and exploit it to deal with the clustering problems of the innovation capability of high-tech industries in China.</p>