Abstract: The rock mass quality surrounding large underground cavern is affected by many factors. The rock classification is relatively complex. Using several rock classification methods synthetically is propitious to accurately determine the integrated category of the surrounding rocks. The classifications called RMR,Q,HC,BQ and RMi are used to classify the rocks of the main workshop on the basis of the experimental data and long-term field geological work. The workshop is the representative cavern of Dagangshan hydropower station. The study is conducted by means of regression analysis. The results show that among Q value, RMR value, and HC value there are exponential relations. The same circs also appears between RMi value and HC value, between RMi value and RMR value. The correlation coefficients for each of the pairs are 0.91, 0.82, 0.87, 0.94 respectively. Meanwhile, Q value and RMi value, RMR value and HC value have shown power relations, linear relations respectively. Accordingly, the correlation coefficients between each of them are 0.90 and 0.85 respectively. Between BQ value and HC value, the RMR value show linear relations, while between BQ value and Q value, the RMi value show exponential relation and logarithmic relation in sequence. The correlation coefficients between each of them are 0.49, 0.38, 0.38, and 0.40 respectively. The results also show that the correlation of RMR,Q,HC, and RMi classification are close related. They can complement and verify each other. They can be effectively applied to engineering practice. The BQ classification has poor correlation with the other four kinds of classification, so it is not very suitable to be used in the power station and similar projects.