Abstract: Aiming at the problem of pipeline failure caused by the development of corrosion defects during long distance pipelines in service, a prediction model of corrosion pipeline failure pressure based on an optimized extreme learning machine was established.The improved SSA was used to optimize the initial parameters of the ELM model, and the adaptive factor based on the Logistic model was introduced to dynamically control the safety value to avoid the premature convergence of the SSA algorithm; the finder-joiner adaptive adjustment strategy was proposed to enhance the local depth mining capability in the later stage of the algorithm; the update calculation method of the finder position was improved to improve the accuracy of the algorithm, and effectively avoid the influence of the randomness of the initial parameters on the prediction accuracy and stability of the model.Taking the blasting experimental data of 61 groups of pipelines with corrosion defects as an example, the ISSA-ELM model was used for simulation calculation.The results show that the average absolute percentage error of the prediction results of the ISSA-ELM model is 1.66%, and the coefficient of determination is 0.9967, which are better than the prediction results of the comparison model.Using ISSA-ELM model as a failure pressure prediction tool for pipelines with corrosion defects has high prediction accuracy and stability, which can provide technical support for pipeline inspection and maintenance.