Abstract: In order to ensure the safe operation of long-distance gas pipeline and to improve the efficiency of gastransportation, periodic pigging operation is required, and particularly, prediction on pig running speed is an importanttechnology for pipeline pigging. As for such shortage of current prediction method for pig running speed as low accuracyand poor applicability, an investigation was made to several domestic pipeline companies for the working flow and successfulexperience of pipeline pigging, and on this basis, the following three prediction methods for pig running speed were summarized: theoretical estimation method, analog simulation method and empirical numerical method. Definitely, the main factors influencingthe running speed of pig in empirical numerical method were analyzed and the optimized empirical numerical method was putforward. By comparing the advantages and disadvantages of the three prediction methods for pig running speed, a complete set ofprediction method of pig running speed was proposed. Finally, taking the pigging operation of a natural gas pipeline as example, the relative error of the three prediction methods was calculated, and the accuracy and feasibility of empirical numerical methodwas verified. The results of study could provide effective guidance to the pigging operation of pipeline operators, so as to enhancethe efficiency and safety of pipeline pigging, and to reduce the cost of investment.