Higher accuracy in the lightning-surge analysis has been pursued; based on circuit analysis. However, the need to analyze phenomena where a plane wave propagation mode cannot be assumed is an outstanding issue. The finite-difference time-domain (FDTD) method is a dominant measure to solve such issue. This study evaluated surge waveforms intruding into a 500 kV air-insulated substation due to a back-flashover through FDTD simulation. The lightning surges were calculated with the inclined angle of the incoming line changed from 0 to 69 degrees and the voltage of the lightning surge intruding into the substation decreased with increasing incline angle of the incoming line. At the open end of the incoming line to the substation at the circuit breaker, the voltage was minimized and maximized when the span length was 20 m and 200 m respectively. This result differed in terms of characteristics from that obtained using the conventional circuit analysis model, whereby the voltage was minimized when the span length was 200 m. The lightning overvoltage at the circuit breaker terminal based on FDTD analysis was lower than that based on circuit analysis. Their relative ratio is 0.557 for a standard incoming line span length of 150 m, showing a 44 % reduction in FDTD analysis from the circuit analytical result. Since the current lightning impulse test voltage is evaluated based on this circuit analytical result, it was clarified that the lightning impulse test voltage level could be potentially reduced with consideration of the inclined conductors at the incoming line span, the vertical conductors of the tower and lightning path, and their mutual coupling with the transmission line. Index Terms -Transmission line, substation, incoming line, lightning surge, backflashover, non-uniform line, lightning protection design.