Abstract: Aiming at the uncertainty of model perturbation and external interference in the unmanned surface vehicle (USV), an improved model-free adaptive control (MFAC) method suitable for the heading control of the unmanned boat is proposed. Theoretical analysis and experimental results show that the heading control subsystem of the unmanned surface vehicle has non-self-balance characteristic, and the direct use of the compact form dynamic linearization based MFAC (CFDL-NFAC) will cause the heading response to oscillate and fail to converge. The historical input items are introduced into the standard control criteria, and a non-increment type CFDL-MFAC (NCFDL-MFAC) is proposed to weaken the integral effect existing in the heading control subsystem of the unmanned surface vehicle. Finally, the effectiveness and practicability of the proposed method are verified through the simulation studies and field experiments of the “Dolphin IB” small USV.