In traditional design of bus sub-station lines, most are based on experience model, which is easy to distribute the same direction lines at different sub-stations, resulting in a decline in passenger travel experience. To solve this problem, this study proposes an automatic calculation method for the optimal bus sub-station lines. Based on the actual travel OD of bus passengers, stopping lines and their departure frequencies, distribution of downstream bus station, and the stopping capacity of each sub-station, a Genetic Algorithm (GA) model is constructed to solve the optimal scheme. In the GA algorithm, the fitness function is “passengers with the same OD board at the same sub-station as much as possible, and travels of the same direction board at the same sub-station as much as possible”, and in the process of selection, crossover and mutation, limitation of each sub-station’s stopping capacity is also fully considered. Finally, some sub-stations in Guangzhou is taken as the test object, with “the total number of passengers boarding at the same sub-station” and other indicators are taken as the test index. Results show that after optimization, the stopping sub-station of bus lines in the same direction are more concentrated, and the number of passengers who fully enjoy taking the line at the same station increases by 71%. Overall, this study can provide engineering application solutions for automatic design of bus sub-station lines.