Many control methods have been designed for the full velocity difference (FVD) model with optimal velocity function, whereas this model can effectively capture the characteristics of vehicular traffic flow by considering both the positive and negative velocity differences. However, the FVD model is required to be linearized for the convenience of designing a typical sliding mode controller (SMC), so that the effectiveness of the controller is only guaranteed in the neighborhood of a steady state. Besides the problem caused by model linearization, the differential operation for obtaining acceleration in the typical SMC amplifies measurement noises. To solve these problems and improve the robustness against complex disturbances, a nonlinear disturbance observer (NDO) based SMC is proposed for the FVD model without reliance on longitudinal acceleration information. The FVD model is considered without model linearization, so that the proposed controller is effective for all of the system states. Moreover, the nonlinear model items, and disturbance caused by wind, are regarded as an “equivalent disturbance,” which is estimated and compensated by NDO. Hence, the robustness against complex disturbances can be improved and the nonlinear model information can be fully utilized. Furthermore, the chattering can be alleviated owing to the NDO based feedforward compensator. Finally, the theoretical analysis and numerical simulations are conducted to verify the effectiveness of the proposed controller.