Abstract: For the prediction of the external corrosion rate of buried pipelines, first of all, a simple analysis of the relevant factors that affect the external corrosion rate of buried pipelines, on this basis, the KPCA algorithm（core principal component analysis algorithm）, CS algorithm（Cuckoo algorithm） and the SVM algorithm（support vector machine algorithm） were introduced, the construction method of the KPCA-CS-SVM algorithm was explained, and the KPCA-CS-SVM model was used to learn the external corrosion rate data and corrosion influencing factor data of 62 sets of pipelines of the actual pipeline.12 sets of data were predicted and verified, and compared with KPCA-PSO-SVM model, KPCA-GA-SVM model and KPCA-LS-SVM model, the advancedness of the model proposed were verified. The results show that the KPCA algorithm can effectively reduce the dimension of the influencing factors of corrosion rate prediction. By analyzing the influencing factors of the external corrosion rate using the KPCA algorithm, it can be found that the redox potential of the soil, the soil resistivity,and the oxygen content in the soil has the greatest impact on the external corrosion rate of the pipeline; the average absolute error of the prediction of the external corrosion rate of the pipeline using the KPCA-CS-SVM model is only 1.89%, the coefficient of determination is 0.9993, and the model training time is only 4.928 s. These three data are better than other models. The research proves that the KPCA-CS-SVM model is a superior algorithm for predicting the external corrosion rate of buried pipelines and can be popularized and applied.