Abstract: Based on the analysis of gas pipeline operating mode and city gas characteristics, using the pessimistic theory and artificial nerve network combination forecast method, a peak-load forecast model on the underground gas storage along a gas pipeline is established and a craft parameter optimization control strategy for the peak shaving of salty cavern underground gas storage is proposed, and a computation process and analysis result from an actual case is produced as well. The conclusion obtained is useful in determining the production plan of salty cavern underground gas storage scientifically and thus obtains economic efficiency.