In order to solve the problems in existing methods of clustering for panel data, grey clustering analysis based on grey relational analysis on accumulation sequences(AGRA) is put. The original data for certain index is accumulated generation firstly and the generation sequences are simulated. Then the dynamic similarity of change trend of the original panel data is characterized by the proximity of generation rate sequences. The AGRA model is comprehensive for deviation, difference and discrete degrees, and a grey clustering analysis is proposed. Based on the AGRA model, the factors in the regional ecological environment system are studied by using the grey clustering analysis. This application is presented to illustrate the effectiveness and practicality of the proposed model.