An incremental nonlinear dynamic inversion (INDI) adaptive control method with neural network is proposed to realize the full-mode flight control of a coaxial high-speed helicopter in presence of model uncertainty. This paper constructed a nonlinear dynamics model of a coaxial high-speed helicopter considering aerodynamic disturbances. Then the INDI control method for attitude inner loop and velocity outer loop according to time scale separation theorem is designed. Furthermore, in order to compensate the model uncertainty and environmental disturbance uncertainty, the BPNN adaptive control algorithm is introduced. Finally, the simulation results are shown to have better robustness by comparing it with the traditional INDI method.