Abstract: Under the influence of environment and pipeline medium, oil and gas pipeline inevitably generate corrosion during operation, which can cause various safety accidents. Therefore, it is necessary to predict the corrosion rate of pipeline.However, there are many factors affecting pipeline corrosion, many factors work together to form an extremely complex corrosion system, and it is difficult to establish a clear functional expression between the corrosion rate and the influencing variables, so the classical corrosion prediction method is difficult to play a role. Therefore, the gray theory was used to study the pipeline corrosion. Firstly, the gray correlation analysis was carried out on the influencing variables and corrosion rate, and the main influencing factors were selected. Then, the grey prediction models GM（0, N） and GM（1, N） were established based on the original data, and the accuracy was tested. The results show that both of them can be applied to the prediction of pipeline corrosion rate, and both of them meet the first level accuracy. Finally, some suggestions are put forward to improve the accuracy of the grey prediction model.