Steel pipeline has been widely used in oil and gas transportation industry because of its unique advantages. However, while providing services, the steel pipeline is easy to be corroded, resulting in volumetric defects, which makes the pipeline thinner and the bearing capacity weaker, further causes the pipeline leakage, and even serious safety accidents. Based on the simplified form of volumetric defects, this paper compares and summarizes the evaluation criteria and evaluation methods, and gives suggestions on the simplified form of defects. The critical factors of single defect and multiple defects on pipeline safety evaluation are sorted out, and the interaction rules of adjacent defects are summarized. Finally, the research direction of multiple defects and the application of machine learning in pipeline safety evaluation are prospected.