Abstract: In order to make full use of limited environmental data to accurately predict oil and gas pipeline corrosion rate, a new type of oil and gas pipeline corrosion rate prediction method was put forward.The EMD experience mode decomposition method was introduced to decompose the environmental data, the limited environmental data were decomposed from its feature which can highlight a new variable to achieve to make full use of environmental data, then, PCA principal component analysis was used to reduce the data dimension of decomposed environmental independent variables to reduce data redundancy.Finally, the GM (1, N) model was improved and its background value was optimized considering the defects of principle and parameters.The modified GM (1, N) model was used to predict the corrosion rate and compared with the predicted results of GM(1, N), EMD-GM(1, N), SVR and EMD-SVR models.The results show that the EMD empirical mode decomposition method is introduced to a certain extent, which can improve the prediction accuracy, based on empirical mode decomposition and modified GM(1, N)model, which is a new type of oil and gas pipeline corrosion rate prediction method, the average relative prediction error is 1.7998%.It has the highest accuracy compared with other models, it is proved that the method has certain advancement.