Abstract: In order to improve the prediction accuracy of corrosion rate of marine pipelines, an optimized deep ridge neural network (DRNN) corrosion prediction model based on hybrid sparrow search algorithm (HSSA) was established.Firstly, through population initialization, the hybrid strategy of improving the position update formula of discoverer, joiner and scout improves the model performance of sparrow search algorithm(SSA).Then use HSSA to iteratively optimize the weights and thresholds of DRNN, reduce the impact of random selection of parameters on model performance, improve model prediction accuracy, and use kernel principal component Analysis (KPCA) to extract the main features in the original data.Finally, it is verified with the data of a submarine pipeline in eastern Hainan, China, which was compared with the other four models. The results show that the prediction results of the HSSA-DRNN model are closed to the actual value than the other models, and the MRE and RMSE indicators are better than the comparison models, which can prove that the proposed model can predict the corrosion rate of submarine pipelines more accurately and reliably.