Natural gas pipelines play a key role in the transmission and distribution of natural gas, and reliable optimization of the pipeline network can lead to safe and guaranteed gas transmission to individuals with minimum risk loss and effective cost. Therefore, we developed an improved layout optimization procedure and applied a new prediction method, the fuzzy comprehensive evaluation method, which is based on entropy weight and backpropagation (BP) neural network (NN). We used it to comprehensively calculate risk loss cost to achieve improved layout during the planning stage and used it to determine the initial risk loss cost. We verified the procedure using a genetic algorithm (GA) for two different layouts, one of which is the improved optimum layout and the other the conventional optimal layout. Finally, the optimized results of four different layouts were determined and the total cost of the new layout is 9.79% and 8.49% less than that of the two conventional layouts. This shows that the new prediction method of risk loss cost provides an efficient and effective means of synchronizing risk loss cost and total cost during the layout planning stage for natural gas transmission network. (C) 2018 Elsevier Ltd. All rights reserved.