To study lane allocation optimization during the ETC promotion period, a data-driven method was presented to predict ETC and MTC volumes in the peak hour of a coming day. Combined with queue theory and drivers’ behavior of choosing toll lanes, a method was proposed to optimize the lane allocation scheme. Actual operation data of 112 toll stations in Shanghai were collected to illustrate the manipulation. The outcome suggests that methods proposed are applicable to predict the daily flow of toll stations, the highest-volume hour of toll station, and the growing trend of ETC traffic, which could contribute to decide the lane allocation scheme.